Examination of witnesses
Jared Sine and Nick Pickles.
Q122 The Chairman: Welcome to our second evidence session this afternoon on our inquiry into the regulation of the internet. We have two witnesses. On the screen, we have Jared Sine, who is general counsel and secretary for Match Group. Thank you very much for joining us, Mr Sine. In a moment we will ask you to say a few words by way of introduction. Our second witness is Nick Pickles, who is the senior public policy strategist for Twitter. Welcome both of you and thank you very much indeed for coming to give evidence to our inquiry. The session will be recorded and a transcript will be taken.
May I ask our witnesses to say a word of introduction and, in so doing, perhaps answer our opening question, which is: has the internet outgrown the regulatory model of self-regulation and co-regulation? Mr Sine, can we start with you with a few words of introduction and a brief answer to the question?
Jared Sine: Sure, I would be happy to. First, I want to thank you, my Lord Chairman and the committee, for the opportunity to present evidence today. Before I proceed with answering the question, I would like to provide a brief introduction to Match Group as well as myself. As was mentioned, I am the general counsel for Match Group and, first and foremost, I want to apologise for not being able to join you in person. I would have loved to be in London at this time of year. Unfortunately, I am unable to do so, but I appreciate your willingness to allow me to join by VTC.
While we may not be the largest of the players that you have spoken with, we believe that as a small to medium-sized platform we offer a unique perspective that should, hopefully, prove informative in connection with your inquiry. It is for that reason we felt it was really important to have someone from our global leadership team join this discussion today.
In terms of who Match Group is, we are an operator of leading online dating brands across the globe, including Match, Tinder, PlentyOfFish, OKCupid and others. We do not take our position as a leader in this category lightly, which is why we strongly agree that public trust in the digital economy is essential and good for all of us. It is for that reason that we support reasonable legislation and regulation in this area, to the extent that it is carefully and thoughtfully crafted, with the input of industry players. One of the reasons for that is because our business model breaks with the traditional business model of online platforms, where platforms effectively offer a free service which is subsidised by the use, sale, monetisation or other licensure of data. Instead, the overwhelming majority of our revenue is derived from subscription services for which our users pay. So already as part of our business model and culture, it is essential that we offer an enjoyable user experience, and one that is safe and engaging for our users, because if we do not, our users will not come back and they will not continue to subscribe, and that would have negative impacts not only on our business model but also on our community at large. As a result, we are constantly innovating and rolling out new feature sets, products, et cetera, to try to ensure that our users have a safe and enjoyable experience on our platforms. In fact, 10% of our workforce is dedicated to safety and moderation efforts.
In terms of your question as it relates to has the internet outgrown the regulation that exists today. As we all know, the internet has grown rapidly and has now penetrated most every aspect of users’ lives. Phones are ubiquitous, online access is ubiquitous and the regulation has, in some respects, lagged behind, which is again why we support regulation or legislation that is consistent with the different needs of the various platforms.
One thing we are concerned about as it relates to regulation is a one-size-fits-all approach that is rigid and targeted at policing and regulating the largest online players. As I mentioned, given the fact that our business model breaks from the business models that are traditionally out there in the online world, it is important that these differences are taken into account, and that any such legislation or regulation is flexible and provides us with the ability to craft solutions to the specific needs and issues that our businesses face.
We also think it is important that as regulation is developed, it is done so in a way that takes into account the various concerns that businesses such as ours have in terms of making sure that the balance of power between various laws and legal regimes is taken into account in connection with such new legislation. For example, in many of our platforms we would love to make further or greater use of automated scanning and other tools that would allow us to review content posted to our platforms, including private messages and other content, to the extent that it contains harmful, offensive or bad behaviour, but, unfortunately, there are a number of privacy law regimes that would restrict our ability to do so. While we would not want access to do so carte blanche, we think that having access and making sure that these regulatory regimes take into account these existing restrictions is critical to ensuring that, to the extent there is new regulation developed, it is easy to adopt and deploy.
We think it also makes sense that any regime adopted or deployed does not end up further shifting the balance of power into the largest players’ hands at the expense of the smaller players, given the fact that we have more limited resources and we have constraints along those lines.
Ultimately, I think that what is most important is that as regulation is thought through and crafted, industry plays a key role in that, so the differences that the various platforms and businesses face can be taken into account in connection with that regulation.
The Chairman: Thank you. Mr Pickles, Could we have a word of introduction from you, please, and your thoughts on whether or not the internet has outgrown the current regulatory regimes?
Nick Pickles: Thank you for the opportunity to appear today. It is a change for me to appear alongside a video camera. This is an incredibly timely inquiry given the public policy conversations we are having around the world. Previously at Twitter, I spent four years based in London leading the company’s public policy work in the UK. I now live and work in San Francisco. My role as a senior strategist is to try to bring together the public policy debates we are having around the world on issues such as terrorism, safety and abuse, disinformation and election integrity, and to bring together the views of our product teams and the Trust & Safety Council, which enforces our rules and writes our policies, to make sure that we try to understand different perspectives when we makes decisions as a company. You may have seen our chief executive testify to Congress last week. We have to be very mindful that the decisions we make have great impacts. We are companies that have real impact in the real world and we need to think carefully, and perhaps in the past we have not thought carefully enough about how we take steps to ensure the health of the conversation that is happening.
The interesting aspect for me when we think about internet regulation is that we often jump straight into the regulation of companies and we miss the regulation of the internet itself. The current challenge in the developing multi-stakeholder model of international internet governance is critical. This is how countries come together through a sea of acronyms, across ICANN, the ITU and the IGF. The reason this multi-stakeholder model is very important is that it has normalised the idea that states cannot control the internet. That means we have a global, free, open internet where information can move across boundaries and borders without states having undue control of that information and the undue restriction of free expression, the erosion of privacy and government censorship. That multi-stakeholder model is far more fragile than people realise. There are concerted efforts from countries around the world that do not share our values to try to get to this state-based internet.
A challenge when looking at national legislation and regulation—and the Information Commissioner illustrated well the existing full spectrum of different laws that apply to companies such as Twitter—is: how can you make sure that something that happens in the UK does not have an effect internationally that changes this very delicate balance? The internet has evolved, driven by free expression, and many services such as ours and the web generally have been driven by self-expression and user-generated content. A lot of the conversations we are going to have today go to the heart of user-generated content, which is people speaking, and are as much constitutional ones as they are commercial, because at the heart of this we are talking about regulating speech. It goes back to the Information Commissioner’s analysis of trying to understand what specific harms we are trying to address, what the policy objective is and what the best way to get there is. One week I am focused on issues around elections and the next week I might be looking at aspects of terrorism. The problems and the solutions are very different. Some are local and some are international and some are between industry and self-regulation.
A real risk we face now is a result of the huge successes that have happened online. The UK has the highest proportion of digital economy of any G20 country and that is as a direct result of the UK’s regulatory framework. At the same time, the Internet Watch Foundation, a remarkable organisation that seeks to remove child exploitation material from the internet, over the past 20 years has reduced the amount of that content hosted in the UK to about 0.3%. It has been tremendously successful in doing that and that is an industry-led and funded body. It is interesting that in its work it finds only about 1% of the content on social media; the rest is elsewhere on the internet.
Today I am looking forward to discussing this notion whereby we sometimes confuse the internet with half a dozen companies, many of which appear quite regularly before Select Committees. The ecosystem is far more diverse and complex. It is a welcome opportunity to have that more complex discussion today.
The Chairman: May I ask our witnesses to keep their answers to questions reasonably concise, as we have quite a lot to get through and we would like an opportunity for discussion with members of the committee?
Lord Gordon of Strathblane: To take a specific example, is there a danger of a conflict between British and American law? In written evidence it was alleged—and I put clearly that it was only alleged—that Twitter had refused to enforce a UK court order without a similar order from an American court. Is that the case?
Nick Pickles: I am not familiar with the specifics of that case but I will happily follow up on it. Absolutely, there are situations where companies find themselves in positions where two different legal jurisdictions will impose different legal obligations. Industry and the UK Government have worked a lot in recent years on situations where UK law enforcement agencies require information that is held by an American telecommunications company—Twitter, Facebook, others—which US law would preclude us from disclosing and where it would be illegal to do something in the UK that is being requested by UK law enforcement agencies. Recently, the CLOUD Act has been passed by Congress and that is specifically to solve this problem between two jurisdictions and, as the Information Commissioner mentioned, will be underpinned by a bilateral agreement between the UK and the US. I will look at the specifics, but we have to be honest and say that one of the big challenges we have as companies is that the overlap of government-to-government legal frameworks is not always easy.
Lord Gordon of Strathblane: I see Mr Sine nodding in agreement. Do you wish to add anything to that answer?
Jared Sine: I would echo the sentiments. I think it can be very difficult for multinational corporations, whether it is Twitter, whether it’s Match Group, et cetera, to make sure that all of our policies and approaches comply with the various legal regimes out there. Unfortunately, there are situations, as previously described, that create these conflicts. We do our very best to address them. The MLAT procedures that used to exist to help facilitate the transfer of data across country lines into the EU, which we would work with EU law enforcement and others who required information in connection with investigations, but it can make it very difficult, for sure.
Q123 The Lord Bishop of Chelmsford: My question, perhaps to you first Mr Sine, is: to what extent should online platforms be liable for the content they host? To add to that, should the liability obligation differ between platforms according to their size or function?
Jared Sine: I believe that platforms should have some responsibility to ensure that they are monitoring and reviewing the content that is on their platforms, but I do believe we have to be very thoughtful in how we craft that liability regime. The reason I say that is, all too often what happens is liability regimes are crafted in such a way that those who don’t monitor or make proactive efforts are actually rewarded in some respects because their lack of knowledge as a result does not impute any kind of liability to them , whereas platforms who are proactively doing the right thing and trying to bring down offensive or inappropriate content are actually held to a higher standard because they are proactively monitoring and reviewing. In that respect, I do believe platforms should be working to monitor, to review, to actively take down content, to the extent that it is reported to them. But in terms of the liability, I do believe there should be good Samaritan or good faith effort-type provisions that allow organisations that are doing the right thing to avoid liability to the extent that, in connection with their efforts, unlawful or inappropriate content slips through cracks of the nets that are established to try to identify that content.
The Lord Bishop of Chelmsford: I will give Mr Pickles an opportunity to answer the same question, but why should the onus be on taking down; why not on not putting up?
Jared Sine: Let us look at our platforms for instance. If you look at the Match Group platforms, we differ quite dramatically from many of the other online platforms. While we do have an area where information is publicly posted—for instance, a person who is looking to find a date or a significant other will post their information in a profile—most of the communication that happens on our platform happens in one-to-one private messaging, and so to the extent you were to require businesses to scan that information prior to it going on to the platform, you would unnecessarily be delaying the ability of users to interact with one another, artificially changing and altering the normal behaviour that people would otherwise engage in and pushing them into platforms that we cannot monitor or help to provide our services on, such as cell phones or other messaging communication tools such as Snapchat, et cetera.
So in terms of creating a system where content is reviewed before it goes up, it would really hamper the experience for the users and the effectiveness of the systems and the tools and the services that we are trying to offer. We do have a number of tools and systems where we scan profiles and other things, on certain of our platforms, as they are coming up, and, on certain of our platforms, very shortly after they come up. We believe we have tools and systems that quickly identify this type of behaviour, but I do believe you would materially alter the services and the efficacy of the business models that are out there online if suddenly there were this up-front moderation that had to take place before any content could be posted to our platforms.
The Lord Bishop of Chelmsford: I may want to come back but of course I know there may not be time. If I have heard you correctly, that could be something to do with the function of the different platforms, so perhaps we should hear Mr Pickles’ views on this one, because Twitter is a very different sort of platform to Match.
Nick Pickles: I think that the rich variety of different services is a point to remember as you are exploring this issue. Twitter has 35 million users around the world, which sounds very large, but in terms of our peer companies, which may have several billion users using multiple services, in some calculations we are quite small. A real challenge with this issue of liability is that—and this is perhaps my Yorkshireman’s scepticism coming through—floated very regularly as the solution to every problem in every shape is to flick this liability switch and all these problems will go away. The Yorkshireman in me says, “Be suspicious of someone who promises you a simple solution because the likelihood is it is far more complex than that”.
It is a question, as the Information Commissioner mentioned, of whether we are talking about illegal content liability, because if you look at things such as Twitter, for example, and the steps that industry has taken around terrorist content, establishing the Global Internet Forum to Counter Terrorism to combat terrorist use of the internet, less than 0.2% of the terrorist content we remove is reported by Governments. We detect the overwhelming majority ourselves. We remove some 75% of the accounts and we take them down before they have tweeted. Again, in the case of illegal content involving child sexual exploitation, a tiny fraction is on social media because social media companies are proactive in addressing the removal of it. Last week, the Home Secretary mentioned in his speech how in one year the National Crime Agency received more than 80,000 pieces of information to assist them to prosecute people.
Baroness Bertin: He also said that you guys had to do more.
Nick Pickles: The interesting point here again is the word “industry”. I am speaking on behalf of Twitter, a company which is relatively mature in this space now and has been party to government conversations, is a member of the Internet Watch Foundation and has worked with the National Center for Missing and Exploited Children in the US for many years, and has resources dedicated to fighting this problem. One of the challenges when we talk about industry is that we are talking about the fact that the Metropolitan Police is currently working with more than 300 platforms to request the removal of terrorist content, when, often, the problem is framed as being one that affects a small number of companies.
To the upload point, a challenge there is that the significant resource constraints on small platforms to do that would have a competitive impact on them.
The Lord Bishop of Chelmsford: Out of interest and as a footnote, do you consider the term “platform” to be an adequate descriptor of who you are? I have to say as a user I find it woefully inadequate. I do not experience you as a platform, not least because of all the adverts you send me. Do you yourself think that is an adequate way of describing yourselves?
Nick Pickles: It is a really good question. One of the challenges in this space is we do not even have the language for some of these things. People try to flit between existing norms and frameworks and putting Match and Twitter in the same bucket is quite challenging in itself. This goes back to what the Information Commissioner was saying, in that perhaps we are approaching this the wrong way. Rather than looking at the internet and platforms as an entire body, we should focus on the specific policy harm. If you take terrorist content, what we see there is challenging because, as the larger companies have dedicated resources to removing the content, it has splintered across the internet into companies based in countries that do not have legal relationships with the UK and that do not engage with law enforcement. It has splintered in a way which was perhaps unintended and unexpected. How you solve that problem now looks very different from how it did three years ago, when there was more of it on the larger platforms. We have taken self-regulatory steps to tackle that problem, and I think we have made real progress. Moira Conway at Dublin University has done a lot of work looking at how the Daesh community have moved and been displaced, and we are very proud of that. It poses a different policy challenge and I think that is not quite there.
The Lord Bishop of Chelmsford: I recognise and acknowledge what you and others have done when it comes to illegal content, but by continuing to define yourself as a platform and therefore a neutral space upon which others stand, it feels to me like it could let you off the hook as regards harmful addictive content which is not necessarily illegal. To use an example from another world, supermarkets have been told, “Don’t put the sweets by the checkout because that is harmful for children’s health and well-being”, but you do still put the sweets by the checkout.
Nick Pickles: One of the challenges of internet policy is that you sometimes have to stretch metaphors somewhat and I am not sure what the sweets are on Twitter, given that you choose to follow an account, you engage with content that you choose to consume and it does not have a calorific value that has a direct read-across to health in that sense. The question mark comes whereby we have rules that cover a broad range of activity. We face a challenge of enforcing our rules in a way that is neutral. You may have seen in the US we have been criticised for applying our rules in a biased way to favour different political groups, and we have been very direct in saying that is categorically not the case. How we enforce our rules is a challenge. The Information Commissioner used the word “trust” and long term, for us, our users trusting our platform is about making sure that the experience they have is healthy and the conversation on Twitter is healthy. That is why we talk a lot about health. It is slightly abstract, but we think there is a different way of looking at content than just: is this tweet good; is this tweet is bad? Context is everything and a tweet can be good in one context and bad in another. We are looking at the health of the conversation across the platform, and are working with the University of Oxford as well as a group led by Leiden University to try to come up with a way of measuring this, because we think there is a longer-term benefit. We see this as not a Twitter-specific issue but as a “health of the public” conversation. We are there to serve that and we hope to share the work we are doing with the whole of industry, so we can try to improve the health of the conversation across the internet and not just one company.
The Lord Bishop of Chelmsford: The Chair is going to shut me up at any moment, but finally and briefly, to give Mr Sine a chance to comment again, since it appeared you would open to thinking of some new language which might be more helpful philosophically in thinking about these new worlds we are in, would you like to think aloud as to what a good word might be other than “platform”?
Nick Pickles: It is a conversation. Industry has to be honest here. This is a conversation we are having as well.
The Lord Bishop of Chelmsford: Would you like us to come up with one?
Nick Pickles: This is where the opportunity is to have a dialogue. I live in San Francisco but I can get on a video conference. We need to have suggestions. Industry is thinking about this, as are academics, and Demos has done some great work on this. One challenge is that we are not going to solve the underlying policy issues around harassment or terrorism by focusing on the definition of the companies involved. Sometimes we put the definitions and the regulatory framework first and it detracts from us saying, “How do we solve this very hard policy question?” Some of that is regulation, a lot of which is already in place, some of it is self-regulation and some of it is societal. One thing that has really struck me in my four and a half to five years working in industry is the belief that you can solve a social challenge by removing content or removing it from particular platforms that are more visible. That is a mind-set that we need to challenge.
The Chairman: Mr Sine, do you have anything to add having had a little more notice of the question?
Jared Sine: In terms of definitions, again, we have used the term “platform” relatively loosely. They are all online services intended to achieve an outcome for the various users. If you look at Twitter, it is a forum for users to have open discussions and dialogues. If you look at our services, they are uniquely structured for users who are 18 years old and older to engage in meaningful relationships and discussions individually and privately. I do not know that we necessarily solve the problem by changing the terminology from “platforms” to “online services”. The key issue remains that while we are online, we are all in many respects doing things very differently at times and other things the same, by virtue of the fact that we are online. I do think that is where the challenge is going to be for additional regulation. Mr Pickles’ idea of, essentially, focusing on the policies and the social harms that we are trying to address is the appropriate way to try to think about it and look at it because, for instance, in terms of offensive content on our platform, the offensive content usually happens between two users during a conversation. We take that seriously. We have a zero-tolerance policy and we remove users to the extent that they are using offensive, illegal or otherwise inappropriate behaviour, whether on or off our platforms, to the extent we can identify it. To legislate for that would be very difficult because you would have to define each type of offensive or inappropriate behaviour. The way to think about it does have to be broader, as Mr Pickles suggests.
The other thing to think about is we work very hard to remove inappropriate content from our platforms; we are by no means perfect at doing it, and there are things that slip through the cracks. However, it is a broader societal question of how do we help society to understand that just because you are online and just because you are potentially hiding behind anonymity, you cannot say whatever you want to say. There are things that I think platforms can do but I also think that there are things that society at large needs to focus on.
Baroness Benjamin: Many young people say they steer away from Twitter because the message is in your face and you do not have a choice whether you want to see it or not, unlike other platforms. Do you think there should be some sort of mechanism for the consumer to decide whether they want to see a message or not? I know on my Twitter account if there is a message I do not want to see, I can block it. Do you think there should be a better mechanism developed so you have a choice as a consumer whether you say, “Yes, I want to see this”—whether it is sexual, abusive or porn, whatever it is; I have as a choice whether I want to see it or not?
Nick Pickles: The committee talked about algorithms in the last session. It is not good enough for industry to make decisions for users. We need to give users more controls that they own themselves. At the heart of Twitter you have the choice of whom you follow. You start with a blank timeline and fill it with people. When I joined the company, if someone replied to you, the only thing you could do was block them. There were no tools, filters or controls. We have rolled out a wide range now so you can say, for example, “Don’t show me tweets that include certain words”. That is often used by people who do not want to see who has won “The Great British Bake Off”, and they mute #bakeoff for a week while they are on holiday. That also allows you to say, “Do you know what, I’ve had enough of the Brexit conversation. I don’t want to see tweets about Brexit”, and you can mute that word. We have rolled out a set of filters that allow people to say, “Don’t show me an account if someone hasn’t confirmed they have a phone number”, and they are using the phone number of their phone—that is a new account; maybe someone who has not changed their profile picture; the brand-new accounts that just tweet at people. By putting controls on to users, we think we can help tackle this. One of the hardest questions in this space centres on the fact that many of the issues are subjective. Trying to write a rule about something that is offensive is very difficult because different people find different things offensive. Where it crosses the line and breaks our rules we want to remove it but where the content does not break our rules, I think you are absolutely right, companies should be doing more. Rather than saying, “The algorithm will sort it out”, they should say, “Here are the controls that you yourself can use”. On Twitter on the home timeline you can choose to turn on or off the algorithm that orders the tweets, so if you want to not have a certain set of tweets shown first, you can turn that off.
The Chairman: You are saying that you develop personal algorithms where as a user I own the algorithm and it is a slave to me rather than to you.
Nick Pickles: Yes, and I think that is one of the things our CEO has been talking about a lot; giving users much more control by using new tools so that you can control your experience and the order in which you see things. By keeping those controls present, rather than just saying we will take the controller and let the algorithm work, we think that builds trust, because if you do not trust the algorithm you can turn it off and see what happens. Those are switches now in your Twitter profile that you can go and look at. As a company we are looking to build much more in the future is that user control to allow you to make sure the choices you make about what you see are the choices for you.
The Chairman: I think we will move on. Lord Gordon.
Q124 Lord Gordon of Strathblane: Staying with Mr Sine, both of you have indicated in the last answer some measures you would take to improve the service you offer what you regard as your communities, even if they are global communities rather than small local ones. Are there any other aspects of the way you try to safeguard and maintain community standards online?
Jared Sine: Sorry, could you repeat the last part of that question? My apologies, the line was a bit garbled.
Lord Gordon of Strathblane: I was recognising that you have already given part of the answer in your previous answer, but how do you establish and maintain community standards for content and for behaviour?
Jared Sine: The way we establish the standards for our community is we first look at the legal requirements that are out there: what is legal, what is lawful content, et cetera. We then take it a step further and try to look at the experience that our users would want to have on our platform. Do they want to be bombarded with solicitations? Do they want to have people treat them rudely or inappropriately? We look at the environment that people would want to operate in and then we try to build our community standards off the back of that. We also make sure that our users agree to those community standards up front. The reason that is important is because then if users violate those standards, we are able to remove them from the platform, to ensure that in that community environment, safe and enjoyable experience, is able to be had by our users. It gets a little grey in our world given the fact these are adults and they are talking often times about intimate things, and what is offensive to one person may not be offensive to another. We try to take a very uniform approach to try to address those types of issues. If somebody is reported multiple times for behaviour that they might think is okay but multiple users think is offensive, we take that into account as well as our community standards. It is a bit of a difficult area for us to operate in given the fact there is not a lot that can be done from a legal standpoint to define what is inappropriate versus appropriate content as it relates to two people trying to engage in a dating relationship.
Lord Gordon of Strathblane: You raised one specific point and mentioned you are different from a normal internet company in that you do not use people’s data. Equally, you must accumulate a vast amount of data on every subscriber you have. Do you do anything with that data? Do you make it available to any other parties?
Jared Sine: We use that data solely to provide our services. We have third parties who help provide our services. For instance, we have relationships with companies such as AWS, to provide technology services, et cetera, to help us offer our solutions on a global basis, but in terms of actual sale or licensure of that data we do none of that. We do use some limited data in terms of advertising on our platforms, but to put that into perspective, it makes up well less than 5% of our total revenue. Again, the vast majority of our service is focused and dedicated on using the data that you provide us to ensure that you have the best experience on our platform, not that that data goes to other platforms.
Lord Gordon of Strathblane: Turning to Mr Pickles, is that what you would like Twitter to do?
Nick Pickles: The first thing is to recognise that it is never static. Sadly, people are very creative sometimes when it comes to being unpleasant to other people. We try to look at what is happening and keep one step ahead of those bad actors. One of the challenges is we change the rules, the bad actors change their behaviour to try to get around the rules, and we are constantly evolving. We established a Trust & Safety Council which brings in safety expert groups and academics from around the world, not just the US and the UK but groups from Korea, Latin America and the Middle East, who share with us, “These are the trends and the challenges we are seeing; how can you make sure your rules are staying ahead of it?”
I would give you one very specific example of the upskirting Bill, which was being discussed in Parliament recently. We wrote our policy some time ago for what has become known as revenge porn, but rather than restricting it to revenge porn, we framed the policy as being non-consensual intimate imagery, which meant that our policy already covered over types of content. Sometimes very specific policy changes are unhelpful because they remove the flexibility to act in other areas. Our Trust & Safety Council is very important. We work with academics and non-profits who are working on these issues, as well as users. Anyone who has used Twitter and seen a new feature being rolled out will know that our users are very vocal in telling us what they think of our changes. There have been occasions where our users have responded and we have undone changes that we have made.
Lord Gordon of Strathblane: Is the very creation of your job which you have taken up over the last few months a recognition by Twitter that it has not been doing enough in this field in the past?
Nick Pickles: Without recreating my job interview, we have a team of 40 public policy people working around the world. Some of those are regional jobs and some are global jobs. Where my role is adding value, hopefully, is in this connection between product and policy and public policy, in that any large organisation always has issues around how it gets different teams to work together and I am, hopefully, bringing different insights to my colleagues in San Francisco. You heard from the Information Commissioner on GDPR and sometimes it is just making sure that people understand what is happening and being that connective tissue. I hope that I am adding value to my colleagues in that endeavour.
Lord Gordon of Strathblane: What processes do you use to ensure that user complaints are dealt with fairly? In a way, that means that the messenger must be identifiable. Is there anything we can do about making sure that only people who have a recognisable address and some method for putting right what they have done wrong are allowed on Twitter? Is there anything you can do?
Nick Pickles: As I mentioned earlier—and the optimist in me likes to remind people—only about 1% of Twitter accounts ever get reported for breaking our rules. The overwhelming majority of Twitter users do not break our rules and do not abuse. We see people using Twitter from other countries, and there was an academic study from Mexico a few months ago which spoke about how the public have higher trust in journalists who do not use their real name when talking about drug cartels because of the physical danger that that job puts them in, so they feel by using their real identity they must not be doing their jobs properly. I have been privileged enough to meet people who practise religions or have political viewpoints that are illegal in some countries and they use Twitter without using their real name to communicate with the world. The ability to be pseudonymous and to not have your real name and real identity on the internet is an important part of enabling free expression. However, we will take action if you break our rules, irrespective of whether you are using your real name or not. We have introduced a system whereby if you break our rules we might challenge you, a bit like a speed awareness course: “You have broken our rules. Here’s the tweet. Before you can come back on Twitter we are going to give you a timeout. You have to stay off the platform for 12 hours or seven days and you have to confirm your phone number so that we know there is a connection between you and your Twitter account”. We have seen that a lot people who get put through that process once do not break our rules again. We think we can improve behaviour and allow people to have their own creative choice about the identity they use online.
Lord Gordon of Strathblane: I am tempted to react by saying that hard cases make bad law. For every case where anonymity is necessary for self-protection, there are, I am sure, dozens where it is simply an abuse.
Nick Pickles: We remove those accounts if they break our rules. We co-operate with law enforcement agencies around the world to investigate those things. There have been court cases in the UK where people have abused people and been prosecuted for it, which is the right thing to do, to hold those people to account. We remove accounts. Law enforcement agencies prosecute people. We think that is the right balance. That should not be at the expense of someone who needs pseudonymity to express themselves.
Baroness Bertin: May I follow up on one point you mentioned earlier, Mr Sine, about being over 18? Age verification strikes me as still a very big problem with Twitter, and, presumably, with Match.com and other more adult content platforms. Could you comment on that very briefly?
Jared Sine: I agree that one of the challenges that all internet companies have is verification of age and identity. We all work hard to try to come up with the best tools and systems in order to do that. Unfortunately, the infrastructure globally is not in place where it is easy to be able to identify and to verify all these things, but we do employ a network of tools and systems to try to make sure that, in the event that anyone who is under the age of 18 tries to get on our platform, they are either blocked or removed very quickly. We start that process by first requiring the user to enter their actual birth date. Of course, a user could add an incorrect birth date and lie about their age. If they do, we have a series of automated as well as manual tools such as photo review, profile review and, in certain instances where it is allowed pursuant to the privacy laws of that country, message review to ensure that we can identify and remove any users who should somehow slip through the cracks who are under the age of 18. It is definitely a challenge that we face. We find in our platforms that are based in the EU, far less than 1% of the accounts created—0.13% of the accounts created—are by underage users and they are removed very quickly thereafter. We grapple with and work hard to try to address that.
Baroness Bertin: We are running out of time.
Nick Pickles: There is a tension, as we have heard from the Information Commissioner, between the objective of reducing the collection of personal data by companies and trying to verify age and identity by using things such as passports and ID cards. Twitter is a service that is overwhelmingly used by older people so we do not see a huge amount of that. Ofcom data looks at that. The issue around age verification and verifying identity depends on whether people have passports or government ID.
Baroness Bertin: If you are an off-licence owner, you are breaking the law if you sell alcohol to children, for example. Should the onus not be put on the companies to make sure they are absolutely not allowing children on to their sites?
Nick Pickles: Are we talking about 18 year-olds, because in the case of alcohol at 18 you have, for example, a passport? Whereas some services allow users under the age of 13, Twitter does not, so how you verify the age of a nine year-old who might not have government ID is much more challenging.
Baroness Bertin: Let us face it, children under the age of 13 are on Twitter and they are taking part in conversations and forums that they are perhaps not mentally able to deal with, and that is where the big problems come with mental health and young children and all the rest of it. It cannot just be dismissed.
Nick Pickles: That is absolutely correct. We are engaging with DCMS on the Green Paper and a wide range of initiatives. Age verification has become seen as the silver bullet to solving a whole range of problems.
Baroness Bertin: You have said that quite a few times and I would pick you up on it. Of course, questions as to who is liable and age verification will not absolutely solve everything but even if they reduce some of these problems by 20%, we are in a better place than we were. I just want to make that point.
Nick Pickles: We all have an opportunity here to work through the consequences of those decisions because the unintended consequences, whether it be restriction in competition or free expression or the diversity of the services that exist, makes this a balancing act, and the work you are doing to understand these challenges is really important.
Q125 Baroness Bertin: Moving on to my question about competition law, obviously, both of your companies buy up rivals and have lots of different names under your brand. Do you think that it is right you are able to do that or do you think there is an issue there with competition law?
Jared Sine: I am happy to address that first. From our perspective, one of the things that we find, based on research, that most users multihome, which means they are using three or more dating apps in connection with their dating activities. Without the ability to offer a variety of different experiences that appeal to different demographics, different tastes and preferences, it becomes very difficult to ensure that you are offering a robust product set. From our perspective, we do not believe that acquiring businesses to provide our user bases with new and interesting services, or services that appeal to their different appetites or tastes and preferences, is something that should be restricted, to the extent that it is not violating the existing competition laws that are out there. We definitely look at those very closely in connection with any acquisitions that we do. I do not believe there should be some blanket prohibition on businesses acquiring other businesses in the same space, given, particularly in our space, the way that the user behaves and operates in the space.
Nick Pickles: We own a company called Periscope and that is quite different from Twitter. We do not own another company that has similarities to Twitter. To add to that, the one thing I would say is there is an interesting tension around industry working together to try to solve some of these problems. The Internet Watch Foundation works on counterterrorism through GIFCT and there is a question mark there around how competition regulators see companies working together to solve these kinds of challenges and decide whether that is anticompetitive behaviour. In counterterrorism work, we are working with smaller companies to help them learn from our experience, but not every small company has the same engagement because they might be less willing to talk to us. That is a question for competition regulators because we want to protect that kind of industry collaboration. The flip side in some of these competition issues is that industry is stronger challenging these by problems working together, and we want to maintain that.
Q126 Lord Gordon of Strathblane: You will have heard the Information Commissioner refer to the powers that her department now has to produce a code of practice about age-appropriate design. What principles do you think should be set out to govern that?
Nick Pickles: As the Information Commissioner rightly said, the UK is ahead of the game on this. Industry is increasingly thinking about it. Our CEO Jack Dawsey said last week that incentives are very important. When you are creating a service, what incentives are you putting there and how do people engage with it? There is a real opportunity for the UK to lead on this. Things such as transparency are important. We heard about making sure people know what they are signing up for. Twitter was rightly held to account for the fact that people felt we were not telling them why we were suspending their accounts. They just got a note saying, “You’ve been suspended”. We now tell you the exact tweet that you have been suspended for and which rule it broke. Transparency builds trust and there should be transparency in both how you are using data and in allowing users to check in real time what you are doing with their data. We now allow people to click through in real time and see, “Which advertising categories have you put me in based on using my data?” Transparency is a key one for me.
Lord Gordon of Strathblane: Does Mr Sine agree?
Jared Sine: In terms of age appropriateness and full disclosure, we had the opportunity to sit down with Baroness Kidron and discuss some of this with her. We think it is a fantastic idea to have more age-appropriate approaches and regulatory regimes, again to the extent that they are tailored appropriately to ensure that, if you are targeting children or those under the age of majority, you are approaching it in a proper way. I agree with Mr Pickles that transparency in that respect is very important and, from our perspective, it is something we laud and we would generally support.
Q127 Lord Allen of Kensington: I think all my points have been covered. My question was on data protection and I was going to ask how you can ensure that people’s data is processed fairly and transparently. I would like to come back to you, Mr Sine. You talked about the fact that you do not provide data to third-party organisations. Do you provide it in an anonymised way if it is not personally given? Secondly, you said that limited data was used for advertising. Can you explain that in a little more detail for me?
Jared Sine: In terms of our data that we use on our platforms, we do not provide data to third parties. In rare instances, we may have a partnership or relationship where we need to use anonymised data. For instance, we recently launched a location feature, and we have a partner who is helping provide location-specific data when someone is at a certain location to show what location they are actually at, so we have to provide some anonymised data to make sure that we get the right information back from that provider, but it is never user-specific information. Again, we share data with third-party providers who are providing services to help us conduct our services, but in general we do not have any need to share anonymised data with third-party organisations for commercial or other purposes that are not related to our services.
In terms of the targeted advertising, like many platforms, we have information as it relates to our users in terms of age categorisation, gender, et cetera, and if a brand comes to us and says, “We want to target this type or category of individual”, we have data that allows us to make sure those advertisements are going to and addressing the appropriate audience, but we do not share that data back with the third party regarding the individuals who saw the ad, et cetera. We may give numbers because they have to understand the impressions and those kinds of things, but we do not provide the user-specific data back to those advertisers. Again, that is a very, very small part of our business.
Lord Allen of Kensington: I think you said 5%. Thank you. Mr Pickles.
Nick Pickles: Several points have already been covered. The interesting question here is that every business nowadays is a data business. It is not just internet companies; personal data is at the heart of pretty much every company now. The first thing about Twitter is we are a public platform which is quite different because people are tweeting publicly to the world their tweets and the overwhelming majority of tweets that are sent are sent publicly. That changes somewhat the nature of the conversation because it is being publicly expressed. For us, a global privacy policy is very important. We do not want a fragmented approach. We want all our users to understand that we have a very clear approach. I am happy to share a copy of our privacy policy because we have done a lot of work focusing on how to make it understandable, how to make it very clear what we are doing, how to make it clear what people can choose to control themselves and how that data happens, and how they can download their own data. Picking up on what the Information Commissioner said earlier, we are also working with peer companies to allow people to take their data to other companies. Data protection is about protecting data, but it is also about empowering users to control their information, and we think if you can move your data to another company and another service that makes for a more competitive open internet.
Lord Allen of Kensington: Which? did a survey recently and it gave us data that showed that 50% of people interviewed did not understand how their data was used. If I asked you how many Twitter customers understood your Ts and Cs and how their data would be used, what percentage would you put on that?
Nick Pickles: That is exactly why we separated our privacy policy from our Ts and Cs, because we wanted it to be a very clear stand-alone document. I used to run a privacy campaign group. I care very deeply about privacy. It is very heartening to see the public debate and conversation that is happening about privacy now. Whether it is about government use of data, corporate use of data, personal control, personal transparency, it is a phenomenonally good thing, and I think it will lead to a more informed public policy debate and savvier consumers. At the heart of all this are businesses, and savvy consumers are the best way of keeping businesses honest.
The Chairman: Can I thank both of our witnesses for the evidence they have given us? Mr Sine, do you have any closing comments or thoughts or anything that you think we might have asked and did not?
Jared Sine: No, I think we have covered the seafront from my perspective. Again, I want to state that from a Match Group perspective we are supportive of regulation in this area, again to the extent that it is carefully crafted, thoughtfully considered, and to the extent that industry is involved. We look forward hopefully to opportunities to continue to work with policy-makers and legislators to ensure that new regulation, to the extent it is developed, is consistent with those principles that we think are important.
The Chairman: Thank you. Mr Pickles.
Nick Pickles: First, may I thank you all for your thoughtful questions? There is a lot of heat on this issue right now and a lot of desire for quick solutions. The big challenge for us all is to think through how to protect free expression and the digital economy and to ensure that consumers are protected. Those three things stand together. I would be delighted to further assist the committee in its work as the inquiry continues.
The Chairman: Thank you again to both our witnesses for their evidence and thank you to those on either side of the Atlantic who made the technology work. We will be producing our report towards the end of the year and your input has been invaluable.
