Science and Technology Committee
Oral evidence: Impact of social media and screen-use on young people’s health, HC 822
Tuesday 16 October 2018
Ordered by the House of Commons to be published on 16 October 2018.
Members present: Norman Lamb (Chair); Vicky Ford; Bill Grant; Darren Jones; Liz Kendall; Damien Moore; Graham Stringer.
Questions 453 - 550
Witnesses
I: Karim Palant, UK Public Policy Manager, Facebook; Claire Lilley, Europe, the Middle East and Africa lead on Child Safety, Google UK; and Sinéad McSweeney, Vice President of Public Policy and Communications for Europe, the Middle East and Africa region, Twitter.
II: Will Gardner, Director, UK Safer Internet Centre; and Sarah Hannafin, Senior Policy Adviser, National Association of Head Teachers.
Written evidence from witnesses:
– National Association of Head Teachers
Witnesses: Karim Palant, Claire Lilley and Sinéad McSweeney.
Q453 Chair: Welcome, all of you. Thank you very much for coming along this morning. In a moment, I shall ask you all to introduce yourselves. May I ask you to make sure that you try to keep your answers as succinct as possible? If you are not being asked a question directly, don’t feel that you all have to answer it. We have quite a lot to get through in the next hour. May we move to introductions? Claire, may I start with you?
Claire Lilley: Good morning. My name is Claire Lilley. I am a public policy and Government relations manager at Google. I am responsible for child safety on our platforms across Europe, the middle east and Africa. Prior to working at Google, I worked for almost a decade at the NSPCC, where I was most recently head of child safety online.
Sinéad McSweeney: Good morning. My name is Sinéad McSweeney. I am vice-president of public policy for Twitter, covering Europe, the middle east and Africa.
Karim Palant: I am Karim Palant, from Facebook, public policy manager with responsibility for the UK.
Q454 Chair: What are the potential harms that young people, in particular, face in using your platforms? What steps have you all taken to seek to address those harms?
Claire Lilley: It is really important to note, first, that there is much about the internet, particularly about YouTube, that is brilliant. It is one of the top platforms that young people use consistently, and they consistently report enjoying their experience on it. They do things such as listen to music, learn how to play the guitar and do their homework. The vast majority of children have a positive experience most of the time. Of course, it is true that sometimes things slip through our safeguards and they are sometimes exposed to content that is not appropriate for their consumption, but we have taken a great many steps to try to ensure that that is not the case.
Q455 Chair: What are the harms that particularly concern you and that you have sought to address?
Claire Lilley: We have very strict community guidelines in relation to a number of areas—the availability of sexual content, the availability of violent content and the availability of hateful or abusive content—and enforce those guidelines aggressively. For example, in relation to sexual content, we do not allow any pornography at all on our platform.
Google has tried to lead the way in making sure that there are ways to safeguard younger children, in particular. For example, we recognise that children under the age of 13 should not be on YouTube, so we have designed a restricted version of YouTube for younger children. It is called YouTube Kids and has a much reduced feature set. You cannot make any comments on it or upload any content. You can turn the search function off completely. It is really a restricted version of YouTube.
The other feature that we feel we have been industry-leading in developing is a tool for parents called Family Link, which enables parents to set the ground rules for their child’s digital use. It is a tool that they can use to control how much screen time their children have and what their bedtimes are set at.
Q456 Chair: Do the parents use it?
Claire Lilley: Yes. I checked last night. On our Play Store, there have been over 50 million downloads so far. Of course, that is also contingent on our having, for example, a multimillion-pound awareness-raising campaign to try to ensure that parents know about it. It is great that these tools exist, and we are really proud of them, but parents need to know about them. We are working really hard to ensure that that is the case.
Sinéad McSweeney: There are a couple of elements to this from Twitter’s perspective. First, you talked about harms. I have been with Twitter for six years. Twitter itself has been on a journey to address the harms and benefits of our platform for all of our users. We have done a huge amount of work to improve the product and our policies, to ensure both that it is a safer place for people and that it is a go-to place for healthy conversation about what is happening in the world.
Q457 Chair: What are the particular harms you are looking at?
Sinéad McSweeney: You talk about young people, but the reality is that young people are not on Twitter, in the main. It is not a platform that is directed towards them, but it is also not a platform that they find attractive. The most recent Ofcom figures indicated that 0% of 12 to 15-year-olds were on Twitter.
That said, we know that we have a responsibility, as an industry player, to contribute to the overall health of the internet space. Ultimately, young people may come on to our platform, and we want to be sure that they come on to it as healthy, good digital citizens. Our focuses range from addressing illegal content and content that is in breach of our own terms of service to surfacing quality content, to ensure that people are not seeing the worst of what people can engage in on the platform.
Karim Palant: Forgive me if I very briefly echo my colleagues in framing the harms that I want to set out in the context of the positives that we believe that the internet generally and social media can provide to young people.
Demos recently published a report that showed that 90% of young people that it surveyed found that social media had a really positive effect on society and their lives. I know that you have heard from witnesses—from YoungMinds and a great many others—who have talked about the really positive benefits.
We would not be here if there were not concerns about harms, and specific harms in particular. There is clearly illegal content that can be circulated on platforms. Things like child sexual exploitation material and terrorist material are obviously a huge priority. We have worked across the industry on building technology to put photo DNA fingerprints, or similar kinds of fingerprints, on images to prevent them from being reshared on platforms and have been building databanks of content such as CSE.
There is a huge range of other kinds of upsetting content that is not as clearly illegal—for example, hate speech or bullying and harassment content. People can report every piece of content on our platform to us. Every image and piece of text can be reported to us.
We have hugely increased the number of people we employ to review that content. We are also building machine learning tools based on the reviews of and replies to particular copies of content. We are building algorithms that will identify that content proactively, in certain cases. Just in the first quarter of this year, we removed almost 3 million pieces of hate speech content from our platform.
A final harm that I know has been raised a lot during these hearings is around the way in which people spend time online and whether that time is valuable and meaningful or can drive negative wellbeing. We take that incredibly seriously. At the end of last year, we published a review both of the research that we had done and of external research looking at the types of interactions that drive positive and negative wellbeing. We have made some significant changes in that area. We may come on to that shortly. I do not want to take too long with my first answer, but we have plenty to add on that.
Q458 Chair: Why do you think there is a disconnect between what comes across as quite strong action that you have all taken to address these harms and many people’s perception that you are not doing enough? People complain to us that it takes far too long for content to be removed or that it is not removed at all. Why is there this disconnect? Do you have any views on that?
Sinéad McSweeney: There are a couple of elements to it. As I said, I have been with Twitter for six years. We would be the first to put our hands up—in fact, we have done so—to say that we did not do enough, particularly in the early years, to address users’ concerns about the ways in which people could report content and about people’s understanding of the rules. There was a lack of clarity and usability.
I have seen a sea change, thankfully, in all that. Our rules are more accessible, it is far easier to report and we are much more transparent about how and when we action reports. We are harnessing technology to ensure that we reduce the burden on the user. All that has changed. As in politics or in media generally, changing a perception very often comes after the work. People are still waiting to see the proof of all those improvements.
Q459 Chair: You think that platforms were probably late in recognising the problems, but you are now doing catch-up. Is that what you are saying, essentially?
Sinéad McSweeney: Twitter and our CEO have said that we did not give enough attention to this at the start.
Q460 Chair: Is that Facebook’s view, as well?
Karim Palant: I absolutely echo a lot of that. Our COO, Sheryl Sandberg, has said something very similar. I have been at Facebook for almost three years. In that time, I have seen a step change in the investment in this area. Earlier this year, we gave some warnings to shareholders that the investment that we were going to make in safety and security would have an impact on our bottom line and would be very significant. That has had an impact.
Last year, we were the first company of this kind to talk about the number of people we have reviewing the content. At the time, the number was around 3,500. We committed to getting to 7,000 by the end of last year. We exceeded that and are well on the way to doubling the number of people who work on safety and security at Facebook from last year to this year. It will go from 10,000 to 20,000. Clearly, this is ramping up very rapidly.
Another issue is playing in some of the feedback you get. A lot of these problems do not manifest only online—they are online and offline problems. An example that I always use is bullying. I know that you have heard evidence on this, but it is very rare for a child—
Q461 Chair: We will come on to that. I take the point you are making.
Karim Palant: The issue with bullying is that if we remove the content and there is a much wider issue, it does not necessarily resolve the issue for the young person, so they will go away with an unsatisfied reaction.
Q462 Chair: I understand that.
Claire Lilley: I will speak to your point about what we are doing to try to reduce the disconnect. There are two things that Google has done this year. First, we are trying to provide people with much better information about what we are doing. In the Government’s internet safety strategy, they have set out that there is likely to be a requirement for a transparency report. Google is already doing that. We launched it earlier this year. We give people very clear information about how many videos are removed, and so on.
More than that, on an individual level, this year we launched a user dashboard. If you make a report, you will now get information about what has happened to that report, which did not happen previously.
Q463 Chair: On a timely basis.
Claire Lilley: Yes. It will tell you what action has been taken in response to your report—whether the content violated a community guideline or a law in a particular country, and on what basis and in which category it did so. It will also tell you what happened to the content—whether it was age gated, whether it was removed, and so on. We hope that this is giving people the reassurance that we are listening to them and reacting when they tell us that there is a problem.
Q464 Chair: Public utilities are always subject to responsibilities and duties. Do you see yourselves as being subject to duties and responsibilities? Do you see it as a question of self-regulation, in effect, or do you think that there is a role for something more than that?
Karim Palant: Mark Zuckerberg, our CEO, has talked about our having a broader set of responsibilities and taking a broader view of what our responsibilities are. That includes being responsible for how people use the platform and what kind of content they post on it. We have set out very clear policies that dictate what we do and do not allow on the platform. We take responsibility both for reviewing reports that users give us—in a timely fashion, as you say—and for building technology and tools for identifying and removing that content, where possible, before it is seen or before it is reported to us. We absolutely take that responsibility.
There are clearly areas where industry-wide self-regulation can play a role. That has been seen around child sexual exploitation material and terrorism, as I highlighted earlier. There may also be areas where Government may well take the view that they can provide co-ordination or a set of rules or frameworks that can impact on how we do things.
Q465 Chair: Would you welcome Government taking that step?
Karim Palant: There are areas, absolutely, where we would say that that could be a positive. For example, this week the Electoral Commission talked about transparency around election advertising online. We have made some moves on that. This week we brought our tools to the UK. They echo similar tools that we have built in the United States. We have done that on a voluntary basis, but that could be a requirement in legislation. I know the Government are consulting on that.
I think we need to take a principle-based approach, where you start with the harm you are trying to tackle. You need to leave room for innovation. We have all talked already about the huge amount of innovation to tackle these harms. Very prescriptive regulation could stand in the way of that, but that does not necessarily mean that there is not a role for Government.
Claire Lilley: The important thing from my point of view is that companies, particularly big companies, do not wait for regulation or legislation before they take action. Google is trying to do that, with YouTube Kids, Family Link and some of the wellbeing tools we have initiated.
I can give you a particular example. About a month ago, we announced the creation of a new tool to identify new, previously unidentified images of child sexual abuse, which we made freely available to all our industry partners and to any NGO that wants to use it. That tool was not brought about by regulation or legislation—it was just doing the right thing, because we have the best engineers and that is where we can really add value in this space.
To answer your particular point, regulation needs to be evidence based.
Q466 Chair: But you do not reject it out of hand.
Claire Lilley: No. As you have heard already from many of the witnesses—academics and clinicians—the evidence just is not there at the minute. That is why we are really happy to see that the chief medical officer will do a review. Legislation and policy making need to be based on that solid evidence; otherwise, there is the danger of having unintended consequences. From my experience working at the NSPCC and ChildLine, I know that very many children engage in social media because it is a lifeline to them. We really need to be circumspect about the limitations that we put on that, potentially.
Q467 Liz Kendall: Will each of you say in turn how many reports of inappropriate content you get each day and how many moderators you have to deal with them?
Claire Lilley: I do not have the exact figure, but I know that it is in the hundreds of thousands of reports every day. In fact, from our transparency report, I know that in quarter 2 of this year we removed 7.7 million videos from YouTube. Similar to Karim, we have made a commitment as a company to increase the number of moderators to 10,000 by the end of this year. We are on track to do that.
Q468 Liz Kendall: How many moderators do you have now?
Claire Lilley: It is in the thousands.
Q469 Chair: Will all of you come back to us with written details of the number of complaints coming in and the number of moderators?
Claire Lilley: I apologise for the fact that I do not have the exact numbers today, but there is large-scale recruitment going on, so any number I gave you today would already be out of date.
Q470 Liz Kendall: But at least we would have the exact numbers. How many moderators you have depends on how many reports you are getting, so we need to know that.
Sinéad McSweeney: Similarly, I am happy to come back with more detail in writing. It is important that we categorise what we are talking about. You used the phrase “inappropriate content”. The kind of content we deal with includes illegal content, which is also against our terms of service, content that may or may not be illegal, but is against our terms of service, and content that is neither illegal nor against our terms of service, but that some people may not want to see. We have various user controls and safe search and quality filters so that people do not have to see that kind of content. That does not mean that people do not want to continue to report it, but the action might be different.
Q471 Liz Kendall: Do you have figures for each of the categories you have provided?
Sinéad McSweeney: We publish figures for terrorist content. We have removed over a million accounts engaged in posting terrorist content.
That is a good segue into your question about moderators.
Q472 Liz Kendall: Will you be able to come back to provide the Committee with the figures we are after? We are trying to assess as best as we can whether, in our view, you have enough people to deal with the problems.
Sinéad McSweeney: I will come back with more detail in writing. As I say, it is coming from a number of different streams. Right now, we have reporting on illegal content and reporting under the EU code of conduct on action on hateful conduct. We are working ourselves, and have made commitments, to update the transparency report that we publish twice a year.
To date, that has related to Government requests. We are looking at ways in which we can add meaningful transparency to that. Meaningful transparency is as important as needing evidence around the impact of social media. Raw numbers, around numbers of reports, are not informative, not least because, as I am sure you have heard before, you have Justin Bieber fans reporting One Direction fans, and vice versa. We need to dig down to make sure that we are giving you meaningful information. That might not be immediately possible, but I will come back with a written response.
Q473 Liz Kendall: How many moderators do you have?
Sinéad McSweeney: That is the point that I was going to make on the terrorist content. Increasingly, we have started to leverage technology to do this work. In the past, we have not released figures around the number of moderators we have.
Q474 Liz Kendall: Why is that? Why would you not let people know how many moderators you have?
Sinéad McSweeney: As we use technology more and more, it is telling only half the story.
Q475 Liz Kendall: You mean that you use algorithms, not people.
Sinéad McSweeney: We use a combination of human moderators and machine learning to surface content for human review. If you look at the 1 million terrorist accounts, 94% of the terrorist content that is removed from the platform is proactively discovered by our machine learning technology. In fact, 74% of it is removed before it even tweets. In order to provide meaningful information, we need to find a way to describe how one balances against the other, so that you are getting the full picture.
Q476 Liz Kendall: Of course, but you would know how many human beings you employ as moderators.
Sinéad McSweeney: Again, it can vary. People use different models to staff their moderation teams.
Q477 Liz Kendall: Could you tell us? Are you able to provide that information to the Committee?
Sinéad McSweeney: I cannot provide that information today, but I can go back and get it for you.
Chair: We fully understand the point you are making about the combination of machine learning and people, but it would be good to get the data on it.
Q478 Liz Kendall: May I ask Facebook the same question?
Karim Palant: The last public confirmation that we had was that, by the end of 2017, we had increased the number from about 3,500 to 8,000. We have made a commitment this year to double overall the teams working on safety and security at Facebook, so that number is changing rapidly and upwards.
Q479 Chair: What jurisdiction do those numbers cover?
Karim Palant: They are global numbers. We are doubling, from 10,000 to 20,000, the total number of people working on safety and security at Facebook this year, so the numbers are changing fast. I will get an update in writing as soon as I can.
That does not necessarily relate directly to the amount of content that we are able to surface and remove. As Sinéad has said, technology is playing a bigger and bigger part in how we do this. To echo Claire’s point about a transparency report, we have published—again, ahead of regulation and before it has been required—a transparency report that sets out the content that we remove. What it shows is that the level at which different types of content are being removed in relation to user reports, versus the amount that is being removed thanks to machine learning surfacing content for human review or blocking it on upload, varies very widely depending on the type of content.
For global terrorist organisations, in excess of 99% of content is removed proactively by our tools. For something like hate speech, where, inevitably, the machine learning will be much harder, because you are dealing with nuanced language issues, context and so on, it is going up rapidly. In the last quarter of last year, about 24% or 25% of the content was removed proactively by our machine learning. That rose to 30-something per cent. in the first quarter of this year. That is a way of testing how well our systems are going. In many cases, the reports that our reviewers review and the responses that they give are what builds the datasets that get that machine learning learned.
Q480 Liz Kendall: You have all touched on this, so I wonder whether you can deal with it as briefly as possible. What do you do when you get a report? How do you prioritise reports? Specifically, do you prioritise reports by age?
Karim Palant: We prioritise by risk of harm. Clearly, age will be a factor, but there is also the type of content. Terrorism content and child sexual exploitation material are at the very top of that. Risk of self-harm is a huge factor. Essentially, we build prioritisation tools, again using AI tooling at the back end that prioritises those cues.
Q481 Liz Kendall: Have you published the way in which you prioritise reports?
Karim Palant: We have set out in a bit of detail in our transparency report an analysis of how we have come to define certain types of content. We talk publicly a fair amount about how we do that. These things will change almost daily, as threats change and as intelligence that is shared either across industry or with law enforcement changes, so I do not know whether we have set out exactly what you are asking for. I can certainly send you what we have published following this meeting.
Sinéad McSweeney: We do not collect dates of birth or other personal information—at least, it is not compulsory. People can supply a date of birth, if they wish, but that comes under our privacy policy and philosophy of minimising the amount of data that we have and hold about our users. Therefore, prioritising by age would not be an option.
Similar to my colleague, we prioritise based on the gravity of what is being reported. Items such as self-harm, threats of violence and so on are prioritised. They are routed to the appropriate team. In recent years, we have built in a lot more transparency about action that we can take and updating people about their reports.
There is one important element around publishing more granular detail about how we prioritise reports. I worked in policing for a long time. We always had a concern around tradecraft and educating people. You would be surprised at the number of people who would jump on that as an opportunity to try to channel their less important report into a queue that they know we prioritise. We would have to balance there.
Claire Lilley: Similarly, we prioritise based on risk of harm. We think that that is the right approach, for two reasons. First, you could have a 45‑year‑old reporting content that depicted an 11-year-old in a harmful situation. We want absolutely to prioritise that. In that case, we think that the age of the reporter is not relevant. Secondly, the reporting systems are designed in such a way as to make sure that people go through the flow and reports get to us as quickly as possible, not to capture additional information at that stage.
The other way in which we do our very best to prioritise child safety measures, in particular, in reports is through a network of what we call trusted flaggers—specialist organisations working in the area. We have over 30 in the UK, including NSPCC ChildLine and the members of the UK Safer Internet Centre.
Q482 Liz Kendall: How much money do you put into trusted flagger organisations? Will each of you say how much financial support you provide to those organisations, if you have used them?
Claire Lilley: Google will give them a payment of $10,000—I will have to check whether it is dollars or pounds, but I believe that it is dollars—a year.
Q483 Liz Kendall: To cover the whole of their work to report to you.
Claire Lilley: As a contribution to their whole system.
Q484 Liz Kendall: It is for the whole of their work.
Claire Lilley: Yes. We do that because the rate of takedowns that we initiate as a result of their flags is much higher. Eighty-eight per cent of what they report will be taken down, compared with an overall rate of 32%.
Q485 Liz Kendall: Isn’t $10,000 a very small amount to act as trusted flagger for the world for Google?
Claire Lilley: There is a whole network of them all around. Quite a lot of them are individual people. Some of them are smaller organisations. Some are larger organisations.
Q486 Chair: Does an individual and an organisation each get $10,000?
Claire Lilley: I believe so, but I will have to check the details of that.
Q487 Chair: Will you come back and clarify it? Sinéad, can you give details of what you pay to trusted flaggers?
Sinéad McSweeney: It is not comparing like with like. We do not have a system of paying people to flag content. We work with many organisations. Again, we have a different demographic of users, and there are different challenges. We work with a number of organisations. Specifically here in the UK, we work with the Samaritans and Childnet. We are working towards anti-bullying week. However, we do not pay people to flag content on the platform.
Q488 Chair: Karim?
Karim Palant: Again, our model is a bit different. We have a wide range of relationships, with a whole range of NGOs, including children’s NGOs, where we fund their work. For example, we offer what we call digital safety ambassadors to every secondary school in the country. We are offering that programme to them over the next year or two. We are working with Childnet and the Diana Award to offer it. Of course, we are providing the resources for that.
We do not have a trusted flagger model per se. What we have is a model for NGOs that work in this space, whether or not they want to have a relationship with us. If they simply want to tell us that they think that we are wrong about things, we provide a channel that enables them to do that. There is no requirement on them to report a certain amount, and no expectation that they will contribute significantly to the removal of a lot of content. It is a direct channel that enables them to raise issues of concern with us.
Often, that is not necessarily about an individual piece of content. It is often about a trend that they may be seeing or about the fact that they disagree with our policies on a particular topic. We will have a direct relationship with them. In many cases, there is a financial arrangement, to do with particular work that they are doing out of that. In many cases, they may not want that relationship, because they think that their independence is important.
Q489 Vicky Ford: My questions are about illegal content. We have heard repeatedly that it takes too long for inappropriate and illegal content to be removed. Why do you think that is?
Claire Lilley: Machine learning really helped us to take down quickly all the videos that we removed in quarter 2 of this year. The vast majority of them were removed very quickly. We have extremely robust systems in order to ensure that illegal content is removed. Let us take my area of specialism, child sexual abuse imagery. Everything on YouTube is scanned to ensure that it is not child sexual abuse. We will look for predatory comments as well. We scan against videos. A couple of years ago, we released a tool called CSA match.
Q490 Vicky Ford: Why does it still take so long?
Claire Lilley: I do not think that it does. In relation to child sexual abuse imagery, it is done very quickly. The Internet Watch Foundation has done fantastic work in the UK. We are very close partners with it and sit on its board.
Q491 Vicky Ford: That is for child sexual abuse. If, come the next election, deepfake videos impersonating each of us pop on YouTube, will you be able to get them down before they go up, as it were?
Claire Lilley: We recognise that on YouTube there are particular areas of content where people need really credible and authoritative information—news, current affairs and politics. We have what we call a top news shelf, or a breaking news shelf, where we prioritise content from an authoritative source and ensure that very good-quality content from the BBC, The New York Times, The Daily Telegraph or whoever it may be, not spammy content, is recommended to people in relation to areas like that.
Q492 Vicky Ford: What about Facebook?
Karim Palant: The length of time to deal with illegal content will vary hugely, depending on the type of content. Through a combination of reviewing and technology, you can get to grips with removing the vast majority of some types of content very quickly. Claire has talked about child sexual exploitation material. Terrorism is another example. As I said, 99% of that material is removed by our tools before it is ever reported to us—much of it within a minute, and often at the point of upload. With that kind of material, of course it is possible.
Where there is a lot more context, nuance and language variation involved—you talk about some newer examples that present new challenges—that is harder. As our investment in AI and so on grows, we hope to get much quicker at proactively removing a lot of that. Of course, as our investment in reviewers, review teams and promoting partnerships with some of the NGOs in this space gets better, we are getting better at removing some of the more nuanced material—the hate speech and so on—more quickly. We have increased the numbers. I think that there is a spectrum.
Q493 Vicky Ford: That does not fill me with a lot of confidence. We are moving to something that, actually, is not particularly new. Quite soon, people will be able to put up a video impersonating someone else, be it a child or a politician, as a fake news story. Will you have the AI to take that down before it goes up?
Karim Palant: I am absolutely sure that people are working incredibly hard to find ways of tackling that issue.
On the point about political fake news, people impersonating politicians and authenticity and transparency, one of the changes that we made in the UK just yesterday is about people needing to pre-approve who they are when they are doing political advertising. There needs to be transparency on the advert itself, saying who has paid for it. There will be an archive of all political adverts going back seven years. Those are transparency tools around who is effectively paying for and running political advertising in the UK that have been introduced here in the UK.
Some of this will be about technological tools to spot and remove particular misleading information. Some of it will be about making sure that users see authoritative, genuine, authentic sources. In truth, we can control only what is on our platform. We cannot control what is happening elsewhere. We need to know that on our platform people are seeing authentic, genuine political communication from the politicians and that they know who they are, that they are based in the UK and that they are playing by the rules.
Similarly, with publishers, we have ensured that the publisher’s logo is shown much more prominently in news feeds. We have made it impossible for people to alter the headline on a news piece when they share it. We are working to bring a range of other tools into that space. Some of it will be about technology blocking that kind of manipulation. Some of it will be about making sure that people see authoritative, legitimate and authentic sources.
Sinéad McSweeney: The two primary concerns under illegal content are terrorist content and child exploitation material. The perception of delays in actioning illegal content is another example of perception lagging behind reality. There is a responsibility on us to be increasingly transparent and more proactive in communicating what we are doing.
I, too, would say that there is not a delay in removing that content. In the case of terrorist content, some of it is removed before it even tweets, with the result that law enforcement is not even referring the content to us, because it is gone before their tools would see it.
There is a lot of learning from the work on both child sexual exploitation and terrorism, from our perspective as a company, about the need to partner with law enforcement agencies—here, the counter-terrorism internet referral unit in the Metropolitan police—to know what trends they are seeing. There is almost a coalition of learning and experience that enables us to identify new trends.
You talk about the next election. I have learned enough at this stage not to predict exactly how the platform might be misused during the next election, but I know that the only way in which we can address that is by continuing to be a learning organisation and continuing to collaborate and partner with other experts. We can even draw on the learnings from hearings like this on the specific area of young people. We cannot be the oracle of both all that is known and all that is right in this area.
Chair: May I make a plea to all of you? We are tight on time. We are running quite a long way behind time, so keep your answers as succinct as possible.
Q494 Vicky Ford: Claire, earlier we heard evidence on the use of YouTube for drill music. This was from someone involved in gangs. They said that the content that you see on the platform is not illegal, but the grooming, as it were, of the young person happens through the chat that then happens in the chatroom. What more can be done about that?
Claire Lilley: We can use machines, where it is black and white. Where it is CSAI or where, clearly, an egregious term is being used—where there can be no debate—we can use machines very effectively to remove that content. When the content is much more context-laden, it becomes a much more nuanced decision, and the content has to be seen by a human moderator. That is where the programme of human moderators comes in.
It is also where the importance of partnerships comes in. It is about working with the Metropolitan police, police services all around the country and our trusted flaggers to understand the nuances of the language that we are seeing—often hyperlocal slang terms, which change very often and which, sometimes, the police are catching up with. We need to bring in a view from the community about what that terminology is and to do so regularly, to keep that up to date.
Q495 Vicky Ford: Germany has now changed the law so that there is a requirement on the platform to remove hate speech and other illegal postings within 24 hours. What proportion of your staff in Europe working on the issue of illegal or suspect content are now working on the German market, versus the rest of Europe? I have been told that it is over 80%.
Karim Palant: That does not sound right to me, but I do not have the figures in front of me.
Q496 Vicky Ford: Can we get them from you?
Karim Palant: A user report anywhere in Europe, whether it be in Germany or elsewhere, gets treated exactly the same. The German law has a requirement for a specific legal flow, which people can raise with us. Our recent figure for the number of people who had reported through that flow was quite low.
Under the German legislation, there is a real risk of requiring companies that do not necessarily have the resources legally to review every piece of content to remove, on a precautionary principle, a huge amount of content that would be perfectly legitimate. It is not a regulatory model that I would say the UK is even looking at, for very understandable reasons, especially given that the UK hate speech laws are far less fitted to that kind of model. Germany has a very prescriptive set of things that are very clearly defined in law as constituting hate speech. That is not the way in which UK hate speech law works, so some of the downsides of the German law would be magnified here.
Q497 Chair: Are you saying that the German law has had no effect on improving the takedown time in Germany, as compared with any other jurisdiction?
Sinéad McSweeney: From our perspective, it is a different system. If people report to us content that is against our rules, various actions are taken. The account may be suspended or put into a locked state until the person deletes that content. For the German law, because you are talking in the main about content that is not in breach of our terms of service and is illegal only in Germany, it is an entirely different process, and a different team. That content is withheld in Germany, so you cannot see it in Germany, but you can see it anywhere else. It is not comparable.
Karim Palant: There is one thing that I would say on the issue of its having no effect. I am not that familiar with the German debate, but in Germany one of the motivating factors behind the law was a perceived rise in hate speech as a whole, particularly around the refugee crisis, I believe. That was seen not just in Germany but more widely across Europe. Has that change in the real world sparked a reaction, in making sure that we have policies and tools in place to deal with that kind of hate speech, in Germany and elsewhere? Absolutely.
Q498 Chair: Are you saying that you do not take down that sort of content any more quickly in Germany than anywhere else?
Karim Palant: Our flow for user review is the same in Germany as it is in the UK.
Q499 Chair: That does not answer the question.
Karim Palant: As Sinéad said, there is an additional flow for content that is illegal in Germany but is not against our terms of service. That is much more common in Germany than it is here.
Q500 Vicky Ford: May I ask you to provide us in writing with what has happened, year on year, on reports of hate speech in Germany, before and after the change in law, showing how much content you are taking down as a percentage of what is being reported and giving comparable data for the UK and other major European countries, so that we have some evidence? Will you also say what resources you are now putting into addressing this issue in Germany, as compared with other European countries?
Karim Palant: I would be very happy to write to you with more details.
Q501 Vicky Ford: I would like, if possible, to have that from all three platforms. Sinéad, I understand that it means that you are stopping viewing of the content only in that one jurisdiction, but it is informative that it is happening. We should be aware of that.
Claire Lilley: I will answer the question from Google’s and YouTube’s point of view, if I may. We have a dedicated user flow to meet the statutes within NetzDG. We have 100 moderators working specifically on that. The figure of 80% that you have heard is vastly overinflated—it is 100 people. It is a new law. What we are seeing so far is that, of all the content that is reported, 30% is actionable. We are seeing a very high rate of content that is not actionable at all. Of that 30%, 93% is removed within 24 hours.
Q502 Vicky Ford: I think that 30% is quite a high number, actually, if it is true hate speech.
Claire Lilley: Yes, it is. I am not saying that that is acceptable at all. What I am saying is that we need to do a better job of educating people about the 22 statutes that come under this, so that we can make it very refined and specific.
Chair: You say that 93% is taken down within 24 hours. What we would like to know, in accordance with Vicky’s request, is how that compares with the period before the law came into force in Germany.
Q503 Vicky Ford: I do not think that it is enough just to say, “The German model is not one that you should follow.” We need really to drill down, to understand whether there have been benefits from that. If three out of 10 reports are true hate speech and are coming down within 24 hours, that is possibly quite helpful. That is just my observation. There should be a bit more follow-up on that.
The National Crime Agency recently stated that tech firms have the technology to be more proactive and could prevent child abuse images from appearing online. Is that true? If so, what is preventing you from doing more? What could we do to help you to do more?
Sinéad McSweeney: The important thing, as a starting point, is that we are leveraging the technology that is available. Colleagues have mentioned photo DNA. There is work going on all the time. We are all part of numerous organisations, both non-governmental and involving law enforcement, that are continuing to push the edges on this.
It is important to remember that less than 1% of the content the Internet Watch Foundation dealt with last year was on platforms like ours. Unfortunately, a lot of this content is elsewhere. We should bear that in mind, both in this area and in the area of terrorism—and, indeed, in other areas. As we know, young people are not on the household names. They are on many other platforms, some of which I have probably never heard of. Increasingly, in areas like illegal content, those of us who have been in the industry a little longer are coming together to help new, younger companies to put robust systems and policies in place. At the end of the day, we all have a responsibility to make the environment safer.
Q504 Chair: Do you think that the National Crime Agency is wrong in saying that you and others could be doing more?
Sinéad McSweeney: When it comes to the safety of children, nobody is ever going to stand up and say that they could not do more to ensure that children are safe, but I think that there are other areas that require more attention. To the extent that we can help those players in the industry, we will do so.
Claire Lilley: At Google, we run our own technology to identify child sexual abuse images. We do that proactively. We have developed technology to identify child sexual abuse video content, and we make that freely available. In the last month, we have developed and released for use by industry players and NGOs content to identify new child sexual abuse material.
Q505 Chair: You think that the charge is unfair.
Claire Lilley: Yes, we do.
Karim Palant: We have talked already about where there are known images. Images that are reported to any one of our platforms go into a databank. We then create a fingerprint for that image and prevent its re‑upload. There are many thousands of images in those databanks. They are automatically prevented from upload, and we scan our platform for those images.
That is what we do for the known images. Clearly, there is a challenge with new imagery. As colleagues have said, the vast majority of such material is not found on platforms like ours. We also proactively seek out and remove nudity from our platform, full stop. That will play a role in this. Something like 20 million images of adult nudity were removed in the first quarter of this year, 95% of which were identified using machine learning.
We have not set out publicly the similar impact around child sexual exploitation material. I know we are in constant dialogue with the NCA and others about that. We are working to be able to share more information around that. Constant iteration and improvement are absolutely necessary. It is perfectly fair for law enforcement to hold our feet to the fire and to say that we could be doing more, and always to be pushing us to do more. We are in constant dialogue with them and are always looking to improve, both as regards what we can remove and as regards what we can report to them and how we can work with them on this.
Claire Lilley: May I make an additional point? Of course, a lot of child sexual abuse is generated by young people themselves, taking what they call sexting images. It is absolutely incumbent on big technology companies, in particular, to do all that they can to remove that content, to keep pushing on that as much as they can and to provide young people with the tools to help them to get the content removed when they know that it has gone beyond their control. But it is also about user behaviour. That is why the education of young people is so important. We need to get this sort of issue on to the statutory curriculum for young people in PSHE and to build in tools like Google’s education programme aimed at eight to 11-year-olds, which talks about exactly this sort of issue and the importance of understanding what you are sharing. Technology is not going to be the silver bullet. We must have a multipronged approach.
Q506 Chair: Can you give us a sentence, Karim?
Karim Palant: On the issue of sexting, revenge porn and things like that, it is very important that we have built a tool—this is for 18-plus—that enables people to provide an image in advance of its being shared online, so that it can have a photo fingerprint and we can prevent it from being reshared. If you feel that you have shared an image with somebody and are worried that they may well reshare it, we can do that proactively. The legal framework around under-18s doing that is very challenging. The Committee may want to consider that.
Q507 Liz Kendall: That is when it starts. It does not start when—
Karim Palant: Absolutely. If young people go to an NGO and try to report themselves as having done that, they may well be criminalising themselves. That is a real challenge in terms of how you build tools. When somebody has done something that they regret, how do we work together? Law enforcement, ourselves and the NSPCC are working on that, to try to find a way through that challenge.
Q508 Chair: This dilemma actually creates an impediment to getting rid of content quickly.
Karim Palant: We built that tool for 18-plus.
Chair: I understand.
Q509 Darren Jones: All of this conversation around transparency of data, reporting, and clarity and understanding of policies, procedures and terms and conditions is day-to-day business for regulated businesses in this country. Would it not just be easier to have a regulator—maybe Ofcom, which has highlighted an interest—to regulate you and all your competitors in this space?
Sinéad McSweeney: There are various elements to that. At the end of the day, a decision on whether to legislate or regulate in this area is one for legislators.
Q510 Darren Jones: That is why we are doing this inquiry. I am asking for your opinion.
Sinéad McSweeney: Exactly. The reality is that, just as we work within the laws that exist today, we will continue to work within the laws that may exist in the future. A former colleague of mine at Twitter is the Australian eSafety Commissioner. We have a really good relationship with her—not just because she is a former colleague, but because we know that her heart and soul is in what she is doing.
Q511 Darren Jones: I understand that. If you do not have a view—if all you are saying is, “We will work with the regulator, if there is one”—that is fine. I am interested in whether anyone wants a regulator to regulate you.
Sinéad McSweeney: The challenge is that if you have different regulators in different countries, applying different standards to what is, in essence, a global service, you may end up impeding the quality and nature of the service and, ultimately, may compromise some of the safety measures that are in existence. That would be my concern. If the decision is made that that is the path we should take, we will have to work within that.
Karim Palant: I thought that Ofcom’s recent intervention was very thoughtful. Our chief executives have talked about the fact that some regulation is inevitable. It has to be principles-based. You have to start by saying what harms you are trying to address. You have to say, “Are we trying to deal with illegal harm and to tackle accountability around that more quickly? Are we saying that there is stuff that Parliament does not necessarily want to ban, but would like to control online?” That is a very difficult and challenging space.
There is then a space around innovation. Each platform is vastly different. The user base is different, the way in which they use the platforms is different, and the way in which the harms manifest themselves is different. The standards that we set ourselves are different, because the user base we are seeking to cater for is different. Any regulator would need to be flexible in the face of that and would need to set expectations, essentially, that responded to it.
That does not necessarily mean that there is not a space for some process. Whether it has a statutory basis or not is an interesting question, but not one on which I would want to express a view. There are clearly areas where we as companies already set goals for ourselves, such as the Global Internet Forum to Counter Terrorism and the Internet Watch Foundation. You could see a role for Parliament and Government in talking to us about those issues.
I think that there is a lot of discussion to be had. I am far from worried about it, in the sense of running away from it, but I think that there is an awful lot of water to go under the bridge. The right way to do it is to start from the harms, to think about how to tackle those, on a platform-by-platform basis, and then to ask where regulation comes in. I think and hope that that is where DCMS is, for example.
Claire Lilley: We should not wait for regulation to take action—we should be taking action already. We feel like we are. Whether we have regulation or not, I am wary of seeing regulation as an endgame in itself. It could be one part of the picture, but it will not account for all human behaviour. We need to do lots of other work around regulation.
My plea would be that if go down the route of regulating, we really need in this space to take into account the views of children and young people, which have not necessarily been taken into account so far. Often their experience differs from what we adults interpret their experience to be. That would be my plea, but we should not wait for regulation before we take action. Regulation also needs to be evidence-based.
Q512 Darren Jones: Thank you for your answers to that question. I have been conscious of the fact that so far in today’s discussion we have focused on the extremes: the things that are illegal and that you are required to remove, such as terrorism and child sexual exploitation. However, for many young people, the day-to-day experience of harm on social media will be bullying, which we have touched on a little, concerns around body image, or mental health issues—things that are not the extremes and are not illegal.
How will the users on your platforms know about and be able to get assistance on those issues, so that they can report on them? I do not want to call them softer issues, because they are not soft, but you will understand my question. I am talking about issues that are not the extremes. How do your tools address that? For example, AI is very good at dealing with images and videos, but not with the nuance of text. How are you taking steps to make that easier for young people?
Karim Palant: On the issue of bullying, Instagram is a very image and photo-first platform; it was initially a photo-sharing platform. One thing that we have done on Instagram, for example, is to develop filters, so that, when people use language that attacks the appearance of Instagram users, they can be blocked at that point.
There are also controls for users, so they can turn off or mute certain comments as well, so people can have controls. There are other filters that we have put on there. Very recently, in the last week or so, we rolled that out to photos and videos, and when people use meme-type text within an image to attack somebody we can build filters to try to prevent some of the stronger comments.
Another thing that is worth highlighting, although I know we are conscious of time, is that one question around young people’s wellbeing online is about how they use platforms. For example, are they using it for passive scrolling, or are they engaged in what we term meaningful social interactions? I find that quite a clunky phrase, but it is the one that we have ended up with.
We published a review of the literature around this just under a year ago and found that positive wellbeing was often driven by meaningful social interactions. The report that Demos recently published showed that. So we changed the way in which the newsfeed worked to prioritise that kind of content over the kind of content that generates passive scrolling. Again, that helps to influence wellbeing.
Q513 Darren Jones: But what about comments or Facebook Messenger messages for those text-based harms?
Karim Palant: I talked about the filters on Instagram. On Facebook, when an under-18 user reports to us that they feel they are being bullied, we treat that bullying as against our terms and conditions, and we will remove it. Context is key, and it is very challenging, as I am sure you can imagine, always to know the context of a particular offline interaction, which means that a comment may look innocent on the page but is actually very upsetting, given the context. One thing we have done is to work with Childline and the NSPCC. When a young person reports bullying to us, we signpost them to Childline to give additional emotional support and offer them that opportunity.
Q514 Darren Jones: And will you report that data in your transparency report on instances of bullying?
Karim Palant: We did not in our first transparency report six months ago, but that transparency report is growing over time and we have another one coming out.
Darren Jones: But will you report on it in your future transparency reports?
Karim Palant: I am not in a position to confirm that, but it will be out in November.
Darren Jones: Okay, thanks.
Sinéad McSweeney: I go back to my original point, which is that this age group is not a big demographic for us. Having said that, we are working to ensure that our rules are more accessible for everybody and our reporting flows are more accessible and intuitive for all our users. That will include young people. As well as over 30 product and policy changes in the last year alone that have made all this territory easier, we are looking at different ways to educate users within the platform. We used to have a very binary “you’re in or you’re out” approach, regardless of the rule that you broke, whereas now there is a graduated response. People are becoming educated around online behaviour as they go through the reporting flow or are the subject of a report.
For me, in my role but also as a parent, what is important when we are talking about mental health and wellbeing is, rather than addressing this subject when we have problems, to take a step back and to build resilience, self-esteem and confidence in young people so that they know how to deal with them. Twitter does a lot of work with that in all jurisdictions, but here in the UK as well. I think that we were the only company that supported the RSPH Scroll Free September. It is about taking the opportunities to engage even when we do not yet have the evidence on how much or little is good or bad; we go with the various initiatives that are happening.
You talk about text-based harms. One thing we have learned is that there is almost more value in addressing yourselves to behaviours rather than words. Context is key, as you mentioned, and sometimes words can be entirely innocuous, but actually there is something going on in an offline environment that makes those phrases more threatening. That is often the case with stalking, for example. There is no one-size-fits-all solution, either technologically, in the product, or indeed in the offline initiatives that we all engage in.
Claire Lilley: We try to make our reporting flow as easy as possible for anyone using the platform. We work with NGOs working with children to understand whether the language is accessible. In my experience of working with young people, I am confident that the vast majority of young people would find the flow easy to use.
We take it to the next level with YouTube Kids where, of course, there are no comments, but children can report a video with one click. We think that that is a good way to serve that need.
To address some of the softer harms that you have talked about, we go to great strides to promote positive content on the platform. For example, we had a lot of content on there around National Coming Out Day, which was designed to support young people who might be unsure of their sexuality and want to explore that on the platform. There was really good quality content—things like body-positive vloggers. We do a lot to support and promote those on the platform.
There is also the issue of comments being context-laden. We put a lot of tools in the power not just of users but of creators to report. We have a particular creator, a beauty vlogger, who has a gap between her teeth. To her, people use the word “gappy” as a huge insult and it is very hurtful to her. To you and I that would not necessarily be offensive or insulting to hear, but it is to her, so she blocks it from ever appearing on her channel. With tools like that we take a really holistic approach to this issue.
Q515 Darren Jones: That is just on YouTube. With Google Search, at a recent event that you held here in the House of Commons I read about your search guidelines for your moderators. I was a bit surprised to see that, for websites that are about self-harm or child sexual exploitation—although CS is not the right example, because I mean the ones that are not illegal—the guideline was to use the moderator to give it a zero rating in terms of quality so that it comes far later in the search results. Why don’t you just remove those rather than putting them on page 15 as opposed to page 1?
Chair: Could all of you give very succinct answers, because we are horribly over time?
Claire Lilley: Let me just check that I have got this right. With content that is not illegal, we—
Darren Jones: You put it on a later page of the search results, so instead of being on page 1 it might be on page 15.
Claire Lilley: I am not overly familiar with this area, but Google is an indexing tool.
Darren Jones: If you are not familiar with it, do not worry, because of the time—you can write to us.
Claire Lilley: Google is an indexing tool that reflects everything that is on the web, and the vast majority of users never click past the first page. Deprioritising it is a very effective tool to ensure that people do not see it.
Q516 Darren Jones: Okay. I had a quick look at your transparency reports while we were asking earlier questions. The Google one, for which the latest day’s entry is for April to June 2018, was very helpful. I was interested that for user and trusted flagger status the UK is tenth in the world in highlighting content. Do you have a view on why the UK is at number 10? Are we not particularly good at reporting ourselves, as users and flaggers?
Claire Lilley: Sorry, is that on the number of reports that were actioned or just the total number of reports that were made?
Darren Jones: It was on the total number of reports.
Claire Lilley: I do not know.
Chair: You can come back to us on that.
Claire Lilley: I can come back to you on it, but the question really is how much of that is actioned. It might be that there is a much higher rate of things being removed as a result.
Chair: An explanation of why we are tenth in that table would be very helpful.
Claire Lilley: Absolutely.
Darren Jones: The data on Twitter were quite old, for July to December 2017. In the UK, at least seven accounts were withheld and only 30 tweets, which seems very low to me.
Sinéad McSweeney: Tweets with illegal content.
Darren Jones: Exactly—but also the other stuff that we have been talking about is not reported and should be reported.
Sinéad McSweeney: We do not yet include it, yes.
Darren Jones: Okay, so in a written response it would be good to know when you will. On Facebook, I noticed that a lot of metrics were still in development and I could not really look at the data. Will you write to us afterwards about what you are going to do to get those up to date, covering areas that are not just about illegal content, and when they will be available?
Chair: Can you all do that?
Karim Palant: Absolutely, yes.
Q517 Damien Moore: I shall be as succinct as possible. The Information Commissioner’s Office has been consulted on the development of an age-appropriate design code for online services that are likely to be accessed by children, potentially including guidance on privacy settings and age-appropriate content. What are the strengths and weaknesses of a safety-by-design approach?
Do you think that there will ever be an age verification process that is robust enough, or do you think that it will always be circumvented by children? Given the fact that we pretty much now just ask for a date of birth or a tick in a box, and children are much more intelligent than that and can get around it, is there anything else that could be put in place, which they could not circumvent?
Karim Palant: On safety by design, safety is the No. 1 priority at Facebook. We take it incredibly seriously with every product that we launch and develop. The way in which our product develops is often in relation to how users use it, sometimes reacting to unpredictable or harmful ways in which people might use it. For example, when Facebook was launched you could not upload photos. That might seem really odd, but we created the photo upload process so that people could put on photos because people kept changing their profile photos. We react to how users use the platform quite a lot, but safety is a big part of that.
On age restrictions, there is simply a tension between how much data you want to keep on 13, 14 or 15-year-olds and how many younger people you wish to restrict from accessing internet products by requiring them to have access to a credit card or a photo ID. The dilemma for policy makers, industry and everybody else is where you draw that line. I do not want to make predictions about the future; there is a very live debate here in the UK about it, and technology and the debate are evolving. I would not make a bold prediction.
Q518 Chair: Are there any jurisdictions where they have taken it further and reached a resolution of this dilemma?
Karim Palant: I am not aware of anywhere where there is an age verification process for people in their teens that would compare with the BBFC process that is still being worked on for 18-year-olds for pornography here in the UK.
Claire Lilley: On the strengths and weaknesses of safety by design, we absolutely think that there are huge strengths to be gained by that—in particular, in relation to young people—which is why Google has put massive resources into developing YouTube Kids and the Family Link app to help parents to manage exactly the sort of things that you have been talking about. It is also why for years we have had a safe search function on our main search engine, which will remove a huge amount—the vast majority—of violent and pornographic content, so that children do not see it. That is why on YouTube we have a restricted mode function, which is often used in education, by schools and so on, but which is also available to normal users such as you and me to ensure that there are tools to make sure that you see only the content that you want to see.
Sinéad McSweeney: I want to cover some of the ground that we have travelled in the last two years in building in a much more robust approach and philosophy around safety into the product itself. Things like having safe search and quality filters on by default, so that people have to actively turn them off, are all part of that.
On the age verification question, again there is a tension between how much data you collect about your users to facilitate it. Of course, children will always seek to do that which they are not supposed to do. I know that in my own house it is about keeping the conversation open, building resilience and ensuring that when somebody encounters something that upsets them or they should not have seen, and the instinct is that they have done something wrong so they are not going to tell anybody, they go to their mum, teacher, uncle or whoever to have the conversation. Building those skills is as important as anything that we build into the sites themselves.
Q519 Damien Moore: So we are building in things to do with content, which is incredibly important, but what about the time spent looking at those pages? Are we building in anything there as well, so that young people cannot look at it as long as an adult can, for example?
Claire Lilley: On the Family Link, parents can set bedtimes and watch times. On YouTube Kids, earlier this year we reduced in response to user feedback the maximum and default time settings from two hours to one hour and from 60 minutes to 30 minutes. We did that in response to what people were telling us—so the tools are absolutely there.
Our chief executive announced at this year’s developer conference—the equivalent of a party conference for tech companies, where they announce their big focus for the year—a whole suite of measures across both our Android operating system and across YouTube to put people in control. It helps them to understand how much time they are using the individual apps for, how many notifications they are getting, how many times they have picked the phone up, giving them a tool to turn the phone over so it will not disturb them or make any notifications. On YouTube, we built a dashboard to tell people how long they spent watching that day, setting prompts so they can say, “After 15 minutes, send me a prompt to ask me whether I want to keep watching.” You have to choose to let it help you to keep watching. Then, of course, on YouTube every single page has a button where you can turn autoplay off.
Sinéad McSweeney: The starting point, and why these hearings are useful, is the understanding of how much is too much, but also focusing on the nature of what people are doing online. The demographic that we have is not in this age group, but some adults may like to be warned that they have been on Twitter for too long. Who knows? As I mentioned, we were the only company in the space who supported Scroll Free September, which probably helped the adults as well as everybody else.
Karim Palant: A few weeks ago, we introduced tools for people to control their experience on Facebook, such as a dashboard to tell you how long you have been on Facebook. You can set yourself alerts on Facebook and Instagram, saying that you want to set a limit and be alerted when you spend beyond a certain amount of time. We have fundamentally changed how the Facebook newsfeed works, so that people have meaningful interactions, which we know drive positive wellbeing, and those are prioritised.
A really big part of this is the conversation between parents and children. I really welcome the conversation that the Secretary of State for Health started in his recent interview. We work with the likes of Internet Matters, the Diana Award and Childnet on digital safety matters. We have produced a parent report to help parents to spark those conversations, which we promote heavily to parents.
Q520 Bill Grant: I have a few brief questions. Colleagues in the Digital, Culture, Media and Sport Committee have recommended that television and radio standards should migrate or be used as a basis for establishing standards for online content. Is that sensible, reasonable and workable?
Karim Palant: Ofcom is the expert on this, and I thought that its recent intervention on this was very thoughtful in drawing the distinction between broadcasting and social media, and the expectations of social media users, including many in this Committee, as to the control over content that we as a platform should exert.
Bill Grant: So that might be a sensible approach.
Karim Palant: What I am saying is that our platforms are very different from broadcast platforms.
Sinéad McSweeney: I agree that they are different platforms and require a different approach.
Q521 Bill Grant: That is very clear. During the morning I have heard mention of other platforms. Google and Facebook are household names, but you mentioned others. Are there any other platforms out there that this Committee or that society should be aware of or concerned about?
Claire Lilley: I direct you to a tool that was developed while I was at the NSPCC, which is called www.net-aware.org.uk, which will tell you the top platforms that young people are using in the UK.
Sinéad McSweeney: Speaking as a mother, gaming is probably one.
Q522 Bill Grant: Finally, from me, as household names do you believe that you have a responsibility for the material that you host on your sites or platforms?
Sinéad McSweeney: A distinction is to be drawn between the legal position in terms of the extent of our liability for the content that is on our platforms, because it is user generated. But as corporate citizens, and citizens of the world, we have a responsibility to ensure, as our CEO defines it, that we are contributing to the health of the public conversation. Twitter is a public platform. From that perspective, we take on a responsibility, whether you call it a moral or a civic responsibility, for the quality and health of the conversation that takes place on our platform.
Q523 Vicky Ford: To pick up on what you said about gaming—I appreciate that you are speaking as a mother, but perhaps the three of you would just comment on this—is that an area where you believe there needs to be considerably more work done on the regulation and child protection issues?
Sinéad McSweeney: I do not think that it is an issue of regulation and child protection. I think that we need to understand it better. I can see enormous benefits. I have an only child, so the extent to which he can be in community with his friends and colleagues is of immense benefit. But we need more research and evidence on the impact; it goes back to that.
Karim Palant: The issue is that none of our platforms are on their own the majority of what young people do online. They will move from one app experience to another, and their screen time will include TV. Other witnesses have told you that—so it is a very mixed picture. I can certainly talk about social media, and we know that young people use social media in a whole wide range of different ways. In fact, we have an event tonight where we are showing a film about six really inspirational young people who have campaigned on refugees. I extend an invitation to all of you to come and watch it. There are lots of things that people do in an inspirational way on social media, and lots of things that people do in gaming that is really innovative and interesting and so on. It is just another way in which young people are using screens and devices. There is a whole mixed picture of what young people do.
Net Aware is a fantastic tool. It talks about all the different apps that people are using. We take our responsibilities very seriously for how people use our apps and platforms. I have talked about the newsfeed changes and the tools that we provide. We do an awful lot of work offline, because that is a really important part of this. For young people, the benefit that they get from being on social media is about how they express themselves offline. Equally, some of the harms that happen offline will come online. That is one of the things that we want to engage in.
Claire Lilley: There is a huge number of really great games out there, which children can learn to develop and create and so on. My daughter uses CBeebies games, and I play some of them with her—and that is great.
I go back to my experience of hearing Childline calls, and there are children who have negative experiences on games, largely when they are playing games that are not appropriate for their age and stage of development, despite very clear PEGI ratings on games. A lot of it is about education of children and parents to make sure that children are playing appropriate games.
Q524 Bill Grant: Could it be an issue that some, though not all, parents are too relaxed about what their children access? Is there a serious need to educate parents?
Claire Lilley: Education is a big part of this.
Bill Grant: I am suggesting that there is a responsibility on the companies and organisations that host the platforms, but by equal measure would that responsibility lie with the parent or guardian of the child?
Sinéad McSweeney: It goes back to the old adage, “It takes a village.” When it comes to children we all have a role to play, and it is a question of how those respective roles can fit together and where we can help—whether it is in supporting Internet Matters in working with parents or supporting Childnet in educating teachers. We will play our part. But you are right—we can put age limits and big signs saying that something is not appropriate, but, at the end of the day, once it goes inside the door, people are allowing their children to sit in bedrooms with devices, with nobody knocking on the door occasionally to see what is going on.
Q525 Chair: May we have a very brief final comment from you, Karim?
Karim Palant: It is about education resource and helping people to start a conversation, and parents being able to talk to kids, knowing that they do not need to worry that they might be ignorant and might not know what children are doing. Our recent survey showed that the vast majority of kids think that parents do not know what they are doing online and do not understand, and they think that it is really positive. Lots of parents agreed. So it is about being able to start that conversation.
I was really encouraged by what the Secretary of State for Health said about some simple guidelines to start the conversation. That is what our parent portal is designed to do. We have recently published a digital literacy library, which is to help teachers to do the same.
Chair: I did say a “brief final comment,” Karim.
Karim Palant: I have finished.
Chair: Good. Thank you all very much indeed. We appreciate your time with us.
Examination of witnesses
Witnesses: Sarah Hannafin and Will Gardner.
Q526 Chair: I am sure that it is a misjudgment, but your entire audience, or almost all of it, has gone—but, welcome. May we have some very quick introductions from each of you?
Sarah Hannafin: I am senior policy adviser at the NAHT, which represents school leaders from early years, primary, secondary and special. Before joining the NAHT, I worked in secondary schools for 18 years as a teaching assistant and teacher, and on the SLT, so I have some experience in the classroom as well.
Will Gardner: I am the CEO of Childnet International, a children’s charity. Our mission is to help to make the internet a great, safe place for children. I am also the director of an EU-funded project called the UK Safer Internet Centre. We are partners with the Internet Watch Foundation, which I believe gave evidence earlier, and the South West Grid for Learning. We are very active as a charity in the education space, working with schools, but also in the policy space, working with Government, industry and others to help to try to shape the environment better for children.
Q527 Chair: Thanks very much. I appreciate your patience—we are starting late. It was a very full-on session. Keep your answers as succinct as you can; do not feel that you both have to answer everything.
Will each of you analyse the benefits and harms of social media use within schools? Are all students equally affected or are there particular issues related to age, background, experience and so on?
Sarah Hannafin: We need to recognise that children and young people see the online world and social media as generally beneficial, essential and really part of their lives. It is an issue for all ages. Our members tell us that the issues are affecting primary school children just as much as secondary. We do not have any evidence that it is not across the age range. Those issues are different. The Children’s Commissioner’s report, “Life in Likes”, flagged in particular the transition between primary and secondary school, and there is a significant change in how young people and children use social media between years 5 and 6 and moving into year 7.
We need to recognise the benefits for children and young people. It is the idea of staying in touch, particularly if they are off school or in the school holidays, or if they are out of school, unwell. They can meet new people and develop interests; there are opportunities for creativity and learning, and all those positive things.
On the flip side, it is so constant that it can be quite overwhelming once you get engaged in the online world. A lot of the risks around contact and being aware that people are not necessarily who they say they are, and of sharing information, have been picked up quite well by young people. That message really seems to be getting across. We need to focus on those things about content and what children and young people should do when they come across content that worries or upsets them.
Q528 Chair: Is there an issue about both the length of time for which children and young people are using social media and the content, or is it more a concern, as you have just indicated, about the content?
Sarah Hannafin: It is about the time as well. The other thing that we are not talking about enough is the potential impact on physical and mental health, which comes with the amount of time, partly, that young people spend online.
Will Gardner: We did a study last February looking at the impact of technology on digital friendships. It will come as no surprise that technology is an important part of young people’s relationships, and the majority of young people would feel isolated if they did not have social media to interact with their peers. We have the great stat that 88% of young people sent a kind message to another young person who was feeling upset. So young people are using the technology to be positive and to be good friends, extending that to the online space.
The risks that we want to talk about are the contact, conduct and content risks. For young people, high in that list would be the content that they come across online, whether it be violent or inappropriate content. Advertising content is another concern for young people. It is also about how their peers are using technology. The flip side of the positive and the friendship is the meanness and those peer interactions that can be negative. On a young person’s list, those things would be near the top.
There are contact issues about adults trying to contact and groom children and young people. We are doing a study at the moment to look at online sexual harassment. There has been much public debate about public figures receiving rape threats via social media, and some 10% of young people in our study have also received sexual threats online. That is a study of 13-year-olds to 17-year-olds.
So, there are some very clear risks. Whether those risks affect young people directly or whether they witness those risks and experience them indirectly, they are very serious. We need to work with young people to find good strategies to respond to the harder-edged stuff about sexual harassment—but also to the other stuff about the pressure within friendships to respond to messages as soon as you have them. The expectation from young people is to respond.
Q529 Chair: Are all the harms in effect extensions of harms that exist offline, or is there something specific and peculiar about social media that introduces new aspects that we need to be concerned about?
Will Gardner: There is a large degree of amplification and facilitation. You can see that in grooming, for example. The groomer might have access to more young people than they would otherwise, in that sense. With cyber-bullying, there is the constant being in touch and the 24/7 nature of it. There is amplification; the audience is bigger and, if there is humiliating content, its life can be longer than it would be if only eye witnesses saw it. So, some elements within that are different, and we need to pick up on them and be able to respond to them.
On the flip side, there is also evidence of what has happened. For teachers and others, it can be enormously powerful for a young person to say, “Look, this is what happened to me,” rather than just saying, “Listen, they did this to me.”
Sarah Hannafin: I would echo that. The internet exposes pupils to risk in a different environment, but a lot of those issues are ones that they would have faced, whether on identity and body image, peer pressure, crime or bullying. But it amplifies them, and it is about how quickly those things can be shared, and that 24-hour culture that sometimes you cannot get away from.
Q530 Chair: When focusing on online issues, is it social media where we should direct most of our attention, or are there other issues, such as online gaming, which we have just been hearing about, which we should be concerned about as well?
Will Gardner: It is about the whole realm. When you think about gaming, you may not appreciate immediately that the same issues of contact risks, content risks and conduct risks—cyber-bullying, grooming and inappropriate content—are just as applicable to the gaming environment as to the broader online environment. The overuse side is also relevant. It is important that we encompass that in the conversations that we are having.
Q531 Darren Jones: We have heard that schools take different approaches to whether pupils are allowed mobile phones in school, in the classroom or at playtime, wherever they see fit. Is there any evidence that you are aware of that shows a link between the use and availability of mobile phones in school and student experience and attainment?
Sarah Hannafin: From our perspective, there is a wide range of mobile phone policies for schools, and that is right. School leaders are in the best position to develop those policies and get buy-in from the pupils, staff and communities to get the right policy in place. Many primary schools among our members have more of a blanket ban on mobile phones.
Q532 Chair: Is that being extended? Are we seeing more and more of them?
Sarah Hannafin: More and more schools are thinking about this issue, realising that they have to consider and make a positive decision about it. They need to think about what their policy is going to be on mobile phones and social media, and why—and make sure that it is communicated.
Q533 Chair: Do you monitor actual bans that are being introduced?
Sarah Hannafin: No. With our secondary members we see much more variation. We see some secondary schools that do impose a blanket ban on mobile phones, but we see a lot of members that manage the use—so phones may be banned at lesson time, but there may be flexibility at break time and lunch time. Some schools also have a policy whereby, even at break and lunch time, phones can be used only outside school buildings.
From talking to those members, we know that those school leaders are talking about how mobile phones and social media are an integral part of everyone’s life. They want to encourage responsible use and create a culture, including in school, about how and when those phones and that content is accessed and used.
A member I spoke to who has a policy like that said that, for him and the schools in his trust, a ban could make things worse. It could push things underground, and mobile phone use might then become secretive and not open, discussed and challenged. It could actually become a bigger problem.
Will Gardner: I do not have evidence, but I shall ask my colleagues and come back to you if we have something in relation to that.
I echo what Sarah was saying. A ban means different things to different people—whether phones are absolutely a prohibited item on school premises, or whether you have a letter from your parents that allows you to have your phone on you, but you are not allowed to use it within the space.
I would agree with trying to encourage positive use. School is a great opportunity for young people to learn how to use technology—not necessarily physically, but responsibly—and about the broader social etiquette, how to stay safe and be respectful to others. There are ways in which to do that within the school community.
It does not necessarily need to be about phones. Some schools are thinking about a “bring your own device to school” policy; when there is amazing technology at home and, perhaps, less amazing technology within school, there may be a way to mobilise that to the benefit of the school. Clearly, there are issues and complications around that, but that is what some schools have been thinking about.
Q534 Darren Jones: Do you think that the centrally produced Government guidelines on mobile phone use, or the maximum amount of time people should spend on social media, which they have been talking about recently, are helpful or a hindrance to schools in being able to make better decisions?
Sarah Hannafin: I think that some evidence-based central guidance on screen time would be fantastic, not just for schools but for families, communities and for all of us.
On mobile phones, it is always good to share practice between schools, and it is much easier to do it centrally, so that schools can see what other schools are doing. In this space, what is important is that the best practice represents a variety of different policies that schools might have approached.
It also addresses the challenges and pitfalls that schools might face when implementing or changing a policy on mobile phones or social media. Sometimes those case studies are so sanitised that it can feel as a school leader that they are not the reality. Seeing what challenges different schools faced and how they acted to overcome them is really useful.
Will Gardner: I agree. The PSHE learning for schools is a really exciting possibility. We wrote, as an example, some guidance for schools on cyber- bullying, prevention and responding to it. We did a survey of educators and a survey of children—a focus group of children—before we put it out. We captured examples of steps that schools had taken practically to prevent cyber-bullying and how to respond to it, and what young people wanted schools to do within that space. Their voice is very powerful.
There is still scope for more evidence and central advice. Last week we were in a primary school, and we asked a group of primary children, “What about screen time? What do you think are the best things that we could be doing within this space?” They were very clear. They wanted to know, “How do we recognise the signs of over-use of technology, and what strategies can we use to deal with those when they arise?” There were very practical responses from young people that could form a part of the solution where we are trying to get that peer-to-peer learning between schools and young people on this issue.
Q535 Chair: Dr Richard Murphy from the London School of Economics has done some research on the impact on students’ GCSE outcomes of banning mobile phones in school and in lessons. He suggested that a ban has a bigger impact on improving outcomes for children who have previously had low attainment. Potentially, there is quite a significant impact on disadvantage. Have either of you looked into that? It is clearly an important consideration that if a ban in class or in school would have a significant impact on improving the life chances, ultimately, of children from more disadvantaged backgrounds, we ought to be considering it.
Sarah Hannafin: We have not looked at that, but I would highlight the idea of what that ban is. We would probably find that the majority of schools did not allow pupils to use mobile phones in the classroom. It is the extent to which those phones might be accepted in other areas of the school.
Will Gardner: I agree. I have nothing to add.
Q536 Liz Kendall: As part of this inquiry, I have been holding sessions in my local primary and secondary schools. The children are really savvy. Many of them understand all of this. They are absolutely all right. The big thing that the young ones—the primary school ones—talk about is people pretending to be someone they are not and not giving any details online.
The main things the older ones are worried about is bullying and body image. Those are the overwhelming things that they have come across. They want to be able to deal with these issues. The question is who is best to help them. What do you think about these issues being included in the review on PSHE?
Sarah Hannafin: Absolutely. Schools have a role to play, in educating young people and giving young people and children a chance to learn and discuss all these issues. Schools are just one part of the solution; many others need to play their role.
The development of relationships—education relationships, sex and health education—is positive. One of the things that is very positive is the creation of some consistency between schools in the experience that children have. The guidance that is currently being consulted on has some very good end-of-primary and end-of-secondary outcomes that focus on online relationships, in particular for relationships education.
Q537 Liz Kendall: What age does this start at?
Sarah Hannafin: Relationships education is primary, and that has a section on online relationships. Relationships in sex education is secondary. It has the same. Health education, again, goes across primary and secondary. The good thing about the health education section, which again has an element of online safety, goes a bit further than relationships and talks about benefits, the impact of screen time and young people learning about the potential impact on physical and mental health.
Q538 Liz Kendall: Does it include anything on people saying horrible things?
Sarah Hannafin: Yes. There is quite a lot about bullying and how to deal with those things. What we really want is children and young people to develop the transferable skills and attitudes that are going to help them make confident and informed choices in whatever situation they find themselves. These subjects are going to help do that. The idea of every child in every school having that entitlement to that learning is really important. We need to make sure that it is resourced properly for schools and there is training for teachers.
Q539 Liz Kendall: That is what I was coming on to. When I visit schools and I talk to teachers, they are way behind. They don’t know half the stuff. They would admit it. So why are we asking teachers to teach kids things that the kids know lots more about than the teachers or parents?
Will Gardner: There are some great ways of managing that. I agree that it can be very intimidating for a teacher to talk to children or teenagers about topics that involve technology, or topics that involve technology and sex can be twice as daunting. There are ways in which we can provide teachers with the resources that are developed in a way that gives them confidence that they can manage them.
PSHE is a really good way of doing that. We have a PSHE toolkit, for example, called Crossing the Line. The young people watch a film and they have to decide, as young people, where the line is crossed between what is appropriate and what is inappropriate behaviour. They are, in effect, working to problem solve. They are working to develop their own rules in a scaffolded environment, which teaching can provide. PSHE lends itself well to having those types of conversations. You are asking the young people to come forward with their own ideas and to have that in a group environment where they can agree what is okay and what is not okay.
Q540 Liz Kendall: What do you think the balance should be? Obviously, we want to remove as many risks as possible to the kids, but the truth is that this is not going away. How much do you think this should be about schools, parents and peers? Do you think overall that we have got to grips with this in the school system?
Sarah Hannafin: Schools are only one part of the solution. To be effective, many players need to play their part, which includes Government, the social media companies, the gaming companies, families, parents and the wider community. To be effective, it has to be all of those people playing the part that they can play.
The pace of development in technology means that adults—parents or teachers—are often left behind. We do not necessarily know what the newest app or social media craze or game is, but we can equip ourselves to make sure that we can talk to young people about the principles of online behaviour—what is good, what is bad, what the risks are—and help them to apply those to the new situations that, of course, are going to come.
Will Gardner: In England we have Ofsted for inspecting schools, and online safety is an element of the inspection. Inspectors check to see what schools are doing in that space.
The computing curriculum has a new safety component within it. Some things are happening already.
We have talked about teachers lacking confidence. The study that we did last December showed that the thing teachers most lack confidence in teaching was online pornography. That is the subject they feel least confident to be able to talk about.
Mobilising young people is a key element. We have a digital leaders programme: 5,000 young people in their school communities are actively reaching out by teaching peers, teaching staff, teaching parents and giving a youth perspective on these issues as they arise.
Q541 Chair: How do these students emerge?
Will Gardner: The schools register for the programme and 10 young people in each school are recruited. They then go to online training and earn badges by delivering sessions for the different audiences. Young people who took on the programme did many things that we had not predicted. They are responding to Government consultations, which we are sending to them, on a range of different topics. They are working with children who have been excluded from school because of their behaviour in relation to technology as they come back to school. There are many avenues for that.
There is also a helpline that I should highlight, the Professionals Online Safety Helpline, which serves professionals, teachers, head teachers, social workers, police and medical professionals. If they have issues around online safety relating to themselves or to their pupils they can contact POSH.
Q542 Bill Grant: As a Committee, we are aware that there is an abundance of advice for parents and teachers around social media and digital literacy. We are also conscious of the varying quality of that advice. Do we have a good understanding of what works when it comes to improving digital literacy and awareness for young people channelled through teachers and parents? How would you comment on the advice?
Sarah Hannafin: One of the issues for schools and teachers in a lot of areas is the wealth of information and advice that is available, the lack of a centralised way of going into that and having a sense that what you are finding is reliable, of high quality and up to date. We have done some work on mental health in trying to improve that journey for schools and teachers.
Online safety is an area that is still quite piecemeal. There is a range of high-quality resources out there, but sometimes it is luck rather than judgment in terms of where you find them. With the move to teaching about online safety with RSE and health education, an opportunity exists for DfE to think about signposting organisations that schools can rely on in producing high-quality, good teaching resources that are regularly reviewed and kept up to date, which for the online world is very important. Rather than expecting the DfE to produce resources, they should take responsibility and identify the organisations and charities that are doing that job very well. That then points schools in the right direction.
Q543 Chair: Do we expect that to happen as part of the process that they are going through?
Sarah Hannafin: In the back of the statutory guidance there is some reference to some resources. We believe that that could be extended a little.
Q544 Chair: You, presumably, have contributed to the consultation.
Sarah Hannafin: Yes. We have not submitted our final response, but, yes, we will be saying that.
Will Gardner: I agree with Sarah’s comments on the statutory guidance, “Keeping Children Safe in Education”. There are connections to organisations, including ourselves, which are providing resources in this space.
There are some things that we know work. You were asking about what works in this space. We organise a Safer Internet Day, for example, which happens every February. We organise it for the UK. It was an EU-wide day that has now become global. It is a celebration by 140 countries. The way we organise the day is that we encourage as many schools to support the day as possible but also as many individuals, MPs, companies, football clubs, charities and different organisations. If we all shout out on the same day, the impact is going to be much greater. We know that last February we reached 45% of children and young people above eight to 17-year olds, and 30% of parents, which for a one-day campaign is quite exciting. We also have good impact stats of those percentages that heard those messages. There are things that we can do.
The school is an important channel for us to try to reach parents and carers. Big awareness campaigns are brilliant. Normally, they are quite expensive. We try to use the conduit of schools—whether that be in book bags, whether schools invite parents in or whether there might be a drama or whatever—a lot when we are trying to get messages out to parents.
Q545 Bill Grant: Would it be that there is a raft of advice, that there are very informative words, but one is piecemeal and one is luck? How do we decide who should introduce the advice? Who should review the advice? Would something like a kitemark or a quality mark give some comfort to some teachers that that is good and up-to-date advice? How do we deal with piecemeal and luck? Surely, that is not what we would wish for in a modern society?
Sarah Hannafin: As I said, it is quite important that the Department for Education takes some responsibility. There are some really good organisations that produce high-quality resources. They have their own quality-assurance process. Things like the PSHE Association or yourselves are reliable. It is about DfE recognising them. That will give schools confidence about where they turn.
Some of the resources that are pointed to keeping children safe in education will be in the RSE guidance. There needs to be some joining together. A fragmented approach is very difficult for schools. The designated safeguarding lead who is engaging with the safeguarding guidance might not necessarily point a link to a curriculum lead. Trying to get some coherence and consistency in the system so there is one place that schools or a teacher can go to is going to be most useful.
Q546 Bill Grant: A single, reliable and up-to-date source.
Sarah Hannafin: Yes.
Will Gardner: I agree. One of the challenges about a kitemark is that things are so fluid. Things change all the time, which is difficult. In some respects, you don’t want to freeze the marketplace as it is. We need to be constantly innovative and challenge ourselves to try different approaches.
We develop a lot of educational material and put it online. It is free for schools to access. The dissemination of that is one of the hardest parts of that piece of work, making sure that the information is in the right hands, in the educational communities across the country. That is a big piece of ongoing work.
Social media is a very useful tool within that because we know, for example, that Twitter is very popular with teachers. When we organise Safer Internet Day our website is one of the first places teachers will come to, but Twitter will be the second place they go to, to try to find educational materials.
We need to try to think strategically what tools we have at our disposal—social media, the Times Educational Supplement or resource centres—to try to help get those resources out.
Q547 Damien Moore: Barnardo’s say they are getting lots of referrals because parents do not know the risks of their children going online at a young age. Are we focusing too much on giving children advice and not enough on giving parents advice?
Will Gardner: We have been focusing on parents, who are a difficult audience to reach. When we started our education programme in schools, maybe 15-plus years ago, we started with primary aged children because that was the age at which they were starting to use technology. We thought we would influence their behaviour as it was in formation. Now, we start with three-year-olds. We do school visits with very young children. At that age, we need to be working to support parents, because the same thing is true. The earlier we engage with parents, the better. This is a conversation that needs to start at the beginning. If parents are supported to set the boundaries in relation to technology use at the start, we believe it is easier for them.
We want them to be engaged at the start. We try different tricks and resources. We have an illustrated picture book that we want parents and children or teachers and children to read together, which will then get that roll going. The key rule that we give to three-year-olds and four-year-olds is: “If you’re not sure online, tell someone.” It is absolutely vital that we support those people whom we are encouraging to tell to make that available, to make sure that you consistently remind young people that you are there to be told and you know what to do if things happen.
Sarah Hannafin: There does need to be more support and advice available for parents and adults generally. Adults are only really starting to become more aware of the impact of time on screen. Scroll Free September happened this year. We are all just becoming aware of what impact it can have on us.
Q548 Chair: Does Government have a role to play?
Sarah Hannafin: I think so. Schools can be a conduit to that wider community and parents, but they cannot be the only answer. There is a broader and more proactive piece of work that Government should be doing to raise awareness throughout. That will affect parents, obviously. Something more needs to be done to make people more aware of the risks, the benefits and, particularly, the stuff about screen time and the potential impact on mental wellbeing.
Q549 Damien Moore: We heard from the Association of School and College Leaders that when they set up e-safety meetings for parents the turnout is particularly low. Can anything else be done to try to influence a higher uptake?
Sarah Hannafin: Unfortunately, it does happen, and it is not just for e-safety. I know of some schools that have had meetings about their sex education programme and had a very small turnout of parents. That is not for the schools’ lack of trying. There is lots of communication and lots of encouragement. Some families are very busy, some choose not to engage and some don’t want to hear that sort of advice from schools.
Schools can do so much and schools will try to get that information out, but we need to look at other avenues. Sometimes, a school website can be useful because a lot of parents might engage with that. If there were particular links or information that will be useful to parents, schools having links on their website might result in parents engaging more. It is a parental choice to go to some of these things. You cannot force parents to engage in a meeting, so we do need to think about other ways in which we can engage people.
Will Gardner: I agree. We run these sessions for parents in schools. Sometimes the turnout is great; sometimes it is less so. There are tricks that we suggest schools could use, maybe around doing a poll of young people to try to highlight what young people are doing on technology and what risks they might be encountering, to try to incentivise parents.
Trying to get young people involved in that process is also a very effective way to try to get parents to come into a school. There is also scope to look outside of school. We try to do that on Safer Internet Day and in our education work. Sometimes we go to big companies and run lunch-and-learn sessions, so at lunchtime we will go in and talk to employees. Parents are, by definition, busy and the time available where they don’t have childcare responsibilities can be, for example, when they are in the workplace. They have an hour for lunch and can come and listen.
We have companies supporting Safer Internet Day who do not have any interaction with children and young people, or much of an online presence, but they have 30,000 employees and they feel it is their responsibility as employers to try to help support the parents in their care. They will help us by sending out messages saying, “Did you know it is Safer Internet Day today? You can find out more information.”
Sarah Hannafin: Picking up on what you said earlier, social media is a good way of communicating with parents. I know from my own experience that seeing some of those things pop up that might be new advice, new information or new evidence is a really good way to get information out to parents.
Q550 Damien Moore: Will, your company set up the first helpline dedicated to helping people who work in children’s areas. How many calls do you receive annually, and which issues keep recurring?
Will Gardner: In the Safer Internet Centre there is a helpline. Childnet is there and a helpline. It is still under the umbrella of the Safer Internet Centre. In the last year, it has had about 800 to 900 contacts from a variety of different professionals working with children—teachers, head teachers, social workers, medical professionals and police; a wide range of professionals—and the issues covered would relate to the young people in their care and the professionals themselves, who might be on the receiving end of bullying by young people or by parents in their school community, for example. There is a wide range of different issues.
Among the different types of issues that they deal with are professional reputation, bullying, harassment, impersonation and online sexual harassment. It can be about the professional’s own professional reputation that is being attacked online.
Chair: Thank you very much. Unless there are any other final questions, that concludes our session. Thank you both for your patience and thank you for the time you have spent with us this morning. It is appreciated.