Examination of Witnesses
Mark Bridge, Matt Reynolds and Alex Hern.
Q152 The Chairman: Can I welcome our witnesses to today’s session of the House of Lords Communications Committee, part of our inquiry into internet regulation? Our witnesses today are experts in the field, and reporters and journalists in the field of technology. We are very grateful to you, gentlemen, for joining us. I know as journalists you are more familiar with asking rather than answering questions, so it is good of you to come along here and share your experience and indeed your expertise with the Committee. Our witnesses are Mark Bridge, who is technology correspondent at the Times; Matt Reynolds, who writes for Wired UK; and Alex Hern, who is technology reporter for the Guardian. It is good of you to be here. We will be recording today’s session. It will be broadcast online and a transcript will be made.
Perhaps I can open by asking you to say a few words about your background and any top-line observations that you might have. In doing so, perhaps you will address our first question area: what do you think are the most serious risks to individuals and society that have been enabled, facilitated or worsened by the internet? How, in your experience, do Government and Governments step up to managing those risks?
Mark Bridge: I am technology correspondent at the Times. I have done that for about two years. I have been at the Times for 10 years, working on other beats. We write quite a lot on these issues, in terms of online harms, issues around regulation and social media issues generally.
In terms of the most serious risks, there are several that stand out. There is the terrorism issue, as in the fact that the internet and social media have allowed terrorists to share their propaganda, to recruit people, radicalise them and incite terror attacks. Almost all of the terrorists behind recent European terror attacks have seen this kind of material. There are also sexual predators who have used the net extensively. Europol had recent figures showing that the amount of this kind of material has proliferated despite all the efforts to crack down on that. Of course, there is misinformation generally and attempts to sow discord in the West and undermine our democracy. There are lots of other potential harms but those are the ones that stand out.
In terms of Governments and how they can address that, it has been a free-for-all until the last couple of years. Not much has been done until a spotlight has been put on to these things and there has been relentless pressure on these companies. It has been a real effort to get them to do anything but the pressure is starting to achieve some results. There are lots of questions about how you regulate effectively. I do not know if there are specific questions.
The Chairman: We will move on to some more specific areas. Thank you, Mr Bridge.
Alex Hern: I have been a technology reporter at the Guardian for five years now. Before that, I covered economics at the New Statesman. I do not disagree with anything Mark has said. There are just a few other areas that I would add. For individuals, a really important risk to take into account is effectively the risk to mental health. The internet and modern communication technologies are radically different from what existed before them and people have a radically different relationship to them. Effectively, the internet flattens relationships. You feel very close to a lot of people. That is broadly new. We as humans are not used to feeling close to 1,000 or 5,000 people. That alone can be extremely difficult and can affect people who are susceptible to mental illness in bad ways.
Also, at heightened moments it can obviously have a negative effect. The stories you hear about online Twitter mobs, for instance, are quite frequently people saying things that would be sort of normal and okay if four or five people said them to you, but we are not able to deal with 10,000 people hurling quite vicious abuse at us all the time, even if, frankly, if three or four people said that to us, we might go, “Hands up. You know what? That was a silly thing to have said”. We are not used to, as humans, having to deal with these numbers.
That same problem affects young people on Instagram feeling like they have a very close personal relationship with 500 influencers. They might think they are close personal friends with 500 beautiful people who work out every day and have wonderful sponsorships—you can see where I am going. The way that you react to these people is different from anything before in society.
Similarly, when it comes to risks to society rather than the individual, everything Mark said is true; but there are also broader problems with society’s ability to even understand and deal with some of the changes that have been wrought. We can talk, using the language we have built up over the last century, of the problems that encrypted messages on WhatsApp pose to policing. We know our society. Our intelligence agencies know how to discuss the problem of people having communications that they do not have access to. We do not really know how to talk and we do not have a language to discuss in the same way, with the same broad participation, the problems of rapid loss of trust in conventional media sources and problems of the flattening and widening of who is trusted. We do not know how to analyse that and so there are clearly risks there. Whether or not it is actually a downside, I do not know. It may prove in the long run that a world in which 10,000 YouTubers provide the bulk of news reporting analysis for the UK is a better world with a better media climate, but it is a risk. That is the sort of risk that is hard to elaborate on without the sort of research that I hope this Committee will do.
Matt Reynolds: I am a reporter for Wired UK. Before that, I covered a similar beat at New Scientist. I get the pleasure of going last so all the clever stuff has been said, but I definitely echo what Alex and Mark were saying, especially Alex’s point on thinking about the flattening of information hierarchies. That seems such an undercurrent to lots of the stuff that is going on today. It is that flattening of hierarchies combined with the fact that you have a very small number of platforms, thinking about Google, Facebook and Twitter. In particular if you think about misinformation, or information more generally, those two combinations seem to be a vastly underlying factor. If you are looking at things like Russian interference in the US in 2016, far-right extremism in the UK, unrest in India on WhatsApp or ethnic cleansing in Myanmar, it seems to me that this is a very obvious undercurrent.
Perhaps that is one side effect of the clustering of so much power and so much of our everyday time in the hands of this very small number of companies. Think how many times today you interacted with Google, Facebook, Apple or whatever. The narrowing of the number of touchpoints that we have represents, like Alex said, a really huge opportunity. In terms of what it enables, it is certainly nothing like we have had before, so it is really exciting, but it seems to me that a lot of the risks stem from that centralisation of power and time spent.
Q153 Baroness Benjamin: From what you have all been describing and what is actually happening, it seems that, in terms of where we are at this point in time, you might say that progress does not always take us forward, because of the everyday events that you described and the behaviour that has been happening. Can I ask each of you: what are the strengths and weaknesses of the current regulatory framework of the internet, which now, as you said, is embedded as part of our lives and our society? How do you feel it can be improved?
Mark Bridge: There is not really a regulatory framework for the internet. It is very hard to achieve one, given its global nature. There are certain areas that are regulated. There are attempts to get the companies to self-regulate, which have achieved certain things. The companies would argue that there are advantages to self-regulation. Very clearly there is a lot of stuff that has been highlighted again and again, and not much has changed or nothing has changed. It is being looked at but age verification, for example, is a huge deal, in that the sites will say we do not have under-13s on our platforms. Ofcom showed that about 23% of 11 and 12 year-olds are in fact on these social media sites that they should not be on. That should be something that can be dealt with fairly easily. It has not been.
Baroness Benjamin: By whom?
Mark Bridge: The Government need to work with companies to develop technical solutions but that should be something that can be done. There should be a way to check the age and identity of those on your service.
Baroness Benjamin: Who should check it?
Mark Bridge: The platforms should, but there should be some standard, presumably, that they would work to.
Baroness Benjamin: You think it should not be a regulatory body; the platforms themselves should check this.
Mark Bridge: Yes. There could be something akin to what is coming in with pornography, for example.
Baroness Benjamin: I see what you mean.
Mark Bridge: Children on these platforms are exposed to so much that is potentially harmful. There was recently a stat on the percentage of children who are targeted by predators; I think it was from an FOI request. Approximately 75% of them had been targeted via the main platforms—things like Facebook, Snapchat and possibly Twitter.
Baroness Benjamin: What you are saying is the BBFC is going to be looking at porn but it should be looking at other things.
Mark Bridge: I do not know whether it should be the BBFC but I think there should be a solution along those lines to keep kids off these platforms. The companies themselves acknowledge that children should not be on these platforms. Again, that can be used with apps that, in theory, are not suitable for 13 year-olds or whatever it is, but six year-olds are downloading them, using them and being exposed to gambling and all sorts of sexual content, whatever it is.
Alex Hern: The strengths and the weaknesses of the internet, when it comes to regulation and when it comes to the internet itself, are the same thing. The strength of the internet is that for most of its history no one has needed to ask for permission to do anything on it and almost anything they can conceive of they can build. That has allowed an incredible flourishing of innovation worldwide. That is not a strength to be dismissed. It is also quite clearly the weakness. The weakness of the internet is that anyone can do anything on it without asking permission. That allows all of the misuses of it that we have seen and heard about so far.
When it comes to obvious areas for improvement, the things I would suggest the Committee looks to are the areas of the internet where that upside—the ability of innovation to flourish—sort of no longer exists. That is because a significant chunk of the internet is actually now quite heavily controlled by a few small power-brokers, which are the same four or five companies that we will hear about for the rest of this session. It is no longer the case that you can do anything without asking permission on a site like Facebook or on a platform like YouTube. These are now fairly centrally controlled. Although Facebook likes to talk about the organic growth that happens on its site, if you actually post something to Facebook, the site will very quickly come back and ask you for money and tell you how many more people you can reach if you pay it. It is acting like a conventional power-broker of the sort that the media industry is used to and that we understand how to deal with, to a certain extent. I would suggest that that also slightly nullifies the upside of a lack of regulation. If it is no longer the case that anyone can do anything, then it is also no longer the case that that light-touch regulation is having much of an upside.
On the specific points that Mark has already brought up relating to young people online, there is a second area where there is again this symmetrical strength and weakness, and that is anonymity online. Again, one of the upsides of anonymity on the internet—the assumption that you do not have to prove you are who you are—is that people can socially and culturally reinvent themselves. They can have communications with people who may not want to talk to them if they knew who they were. They can, for instance, reach out to journalists without revealing their identities, to leak information or just to provide expert insight. That is an upside; and again, it comes with its symmetrical downside. I do not think it is easy. I do not think you can simply remove anonymity from everyone, which is what age restrictions would entail because you obviously cannot ask only young people to prove they are who they are; you have to ask everyone or you ask no one. There is a strong downside to that.
Again, the places I would say it fits most obviously with the current state of the internet are the places where that anonymity has already been removed. Facebook enforces a real-name policy. If people have names that the site moderators do not recognise as real, it asks them to upload identification—typically state identification. It has already done that in some unpleasantly heavy-handed ways. Native Americans in America have reported being disproportionately targeted by this because traditional native American names sound, to someone who has not heard them, like they might be a joke name. A lack of cultural awareness has led to unpleasant enforcement.
Similarly, people with famous names, frankly, have reported being unfairly targeted. If your name is Mark Zuckerberg and you are not the Mark Zuckerberg, I cannot imagine you get to use Facebook without providing some ID. I do not think there is as much of a downside on those platforms to removing that presumption of anonymity as there would be if we tried to apply the same rule to the internet at large. It is that symmetry. The benefits of light-touch regulation are quite broad and quite intrinsic to the internet, and I would ask the Committee to be careful when considering whether to remove them and to look at whether or not they are in practice still there in the first place.
Matt Reynolds: I would agree with a lot of that. The light-touch regulation of the internet is essentially what has shaped it as an incredibly useful resource now. Essentially it means that what we have ended up with is a fractured, self-regulatory environment where what passes as hate speech on Facebook is not the same as what passes as hate speech on Twitter, and the same with YouTube. This is really problematic. At the moment you get a very ad hoc approach to dealing with any of these issues. Take the banning of Alex Jones. I think Spotify removed a podcast, then you see it in Twitter and then you see it in Facebook. Actually, if he was in violation of all of their terms of services or community guidelines at the same time, there should have never been any question that you would have this conga line, if you like, of banning.
It suggests to me that there is a space for at least some kind of alignment between these. If this is where we have our public conversations, if this is where we think about it, we need to be thinking about what kind of standards we can enforce and what is expected. At the minute, it is very responsive to press and media coverage but it does not have any coherent outline. I see that when Germany basically forced Facebook and social media to do more to combat hate speech, Human Rights Watch basically attacked that as a worry for overt censorship. I really appreciate that it is a very difficult thing but some sense of alignment on that is a good idea.
To your point on thinking about whether you offload this responsibility on to a regulator or whether you ask the platforms to self-regulate, we have to expect that at this scale that we are talking about, with the scale that these platforms are, they have to build self-regulation into their very scale. I do not accept the argument that you get to have this reach of 2.2 billion people and then say, “We forgot how to check out like we did before”. It should be a requirement of building these platforms and building that reach that you have the ability to be on top of that.
“Self-regulation” is the wrong term; self-moderation that is overseen by a regulator makes a lot of sense to me. We should not offload the responsibility of catching it to the press or to whoever, but we should be realistic about who is capable of keeping that eye out. It seems to me that it is going to be the platforms that built these things in the first place.
Baroness Benjamin: What about current regulatory bodies such as the ICO? Do you think that they are effective and properly resourced in regulating the internet and could solve the problems that we are facing?
Matt Reynolds: My sense is no, insofar as thinking about the ICO specifically in terms of the amount of money they are able to leverage for fines. It is hard to see, because it very early days in terms of enforcing GDPR, whether companies respond to things like that. I do not have an awful lot of faith that they are well enough resourced or that their remit is outlined clearly enough or that their ability to effect change in those organisations is realised or felt enough.
It feels to me that possibly—and this is no surprise because of the headlines it has gathered—the EU’s competition authority has probably done the most in making tech authorities listen. This is probably for a couple of reasons: first, just by virtue of the amounts they can leverage and force; secondly, if you are Google, you are thinking about what affects a market of 300 million people and how you can shape your service so it aligns with those. This is a point we will get on to, but the weight of having policy that aligns with a broader environment like the EU makes an awful lot of sense. It gives it a lot of persuasive power.
I am not totally convinced. Aside from the ICO, in the UK it is so fragmented. Where was the Advertising Standards Authority when we were talking about Facebook ads? Consistently, existing bodies have not stepped up or seen that their remit extends to the online world, if you want, which is literally just our real world. So far, there have been quite a lot of failings in that respect.
Baroness Benjamin: Do you think there should be perhaps a global regulatory framework body that gets it all together?
Matt Reynolds: Do you mean internationally?
Baroness Benjamin: Yes. At the moment, you are saying it is all fragmented, people think it is somebody else’s problem and we are not really having that joined-up, holistic policy that we should have, because, as I said, we are progressing and progress does not always take us forward. How much further are we going to progress without literally making sure that we have a framework that is looking after the interests of people, especially our children?
Matt Reynolds: The worry with a global authority is that you think that there are a billion people living in China who have a very different idea of privacy than we do. There are so many people in America that again have a very different idea of privacy and rights than we do. There is a pay-off, right? If you are Google, a site of 300 million or a consumer pool of 300 million is sizable; it would influence your policy and it would make you think.
I am not completely against the idea but my worry is these voices in Silicon Valley are so influential that when you bring in factors like China, perhaps finding this perfect point that everyone works towards just might not exist. What can we do in the UK? We can probably defend the things that we think are right for the internet, and we should not be that afraid to try to enforce that as best as we can and say that if the global standard does not step up to what we think is appropriate and the type of things that we want to see, that is something you have to pull in line with and that we do not have to step back towards. That is probably impractical but I would be a bit worried about the kind of influence a global body would have.
Q154 Lord Gordon of Strathblane: I have a further point. It is almost inevitable that however well sourced regulators might be, either present ones or possible future ones, at some point they are going to recommend to Government that somebody legislates to do something. Do you think Government, as we have it at the moment, and Parliament are properly structured to cope with something as fast-moving as the internet? If, as I suspect, you think it is not, how would you recommend it improves?
Alex Hern: The fast-moving nature of the internet is not as antithetical to regulation as it might seem. There are a lot of regulations and a lot of proposals that can be phrased in human-readable language, rather than technical language. I agree it would not be appropriate to pass legislation saying that, for instance, if something has more than 5,000 retweets it should be deemed that the platform should pay attention to it because it has been seen by a lot of people. That sort of thing would never work but with the Advertising Standards Authority, for instance, it is not clear to me that you need to write technical language into their code of practice for their remit to be extended to the internet. They judge advertisements based on non-technical standards: accuracy, truth. Those have not been changed that much by the internet. I do not quite understand why something being online means that the legislation has to be written in such a way that it cannot be future-proofed.
Q155 Baroness McIntosh of Hudnall: I want to ask any of you who want to pick it up about other regimes. You made the point, Mr Reynolds, that we ought to be looking at what we can achieve here given that attempting some kind of global reach is probably beyond any particular parliamentary competence. What is your view of, for example, what has happened in Germany, where they have politically taken a much tougher line on certain kinds of regulation than, for example, we yet have? Do you think that has been a move in the right direction, or do you think it has had unintended consequences in terms of free speech or any of the other things it might have impacted on?
Matt Reynolds: It is a really difficult question, which is no surprise. I know this sounds maybe like an evasive answer but it comes down to the fact that if the German Parliament is tasked with enforcing policy that it believes is right to protect people in Germany and is right for that, I do not necessarily see that that would be an example that we would want to follow just because they could and because they have. It underlines the idea that we should be having this public debate. An organisation like Doteveryone, which I presume spoke to you as part of this process, is really good for having that debate and deciding what we think is acceptable and what should be legislated for and what should not. There is a lot of work assessing where we are at with that conversation that we need to come to before we think about how we can put in place laws that align with this.
I would also add to what Alex just said. If you think perhaps about our free speech law or our speech laws more generally, I am interested to know what specifically about them being online means we have to update or do something new. We should be thinking, “If these are the standards that we apply to public speech, public broadcasting or that type of thing, might we think usefully how we apply it online?” I do not think we are coming completely from scratch, not saying, “Should hate speech be on the internet?” so much as, “What is acceptable to come across in your daily life?” Perhaps if we think about that, we have a body of legislation, a body of evidence and clever people who are thinking about this, which we can leverage more easily.
Baroness McIntosh of Hudnall: You might think that the difficulty there is actually in implementing and sanctioning. The basic principles may be right but applying them in the world of digital technology is perhaps slightly more difficult, would you agree?
Matt Reynolds: I agree that it is more difficult. To slightly push back on that, I completely agree that it is difficult and the scale makes it difficult, but it being difficult should not be an excuse. When we are talking about the scale and difficulty, we are also talking about companies that have vast resources, are hugely influential and are in a position where, frankly, they should be stepping up to it and having this conversation. A factory manufacturer does not scale up and say, “It is unsafe. We are moving really fast. We are trying to get to everyone and do loads of products”. We need to start, in our terminology and how we think about this, to demand a lot more and expect a lot more, because they are huge, they are influential and they are already reaping the benefits of scale. It is time to think about how you apply the responsibilities of that. Difficulty cannot be an excuse, although I completely appreciate that it is difficult.
Q156 Lord Goodlad: You have already covered some of this in your discussion following Baroness Benjamin’s original question, but could I ask how effective you think platforms are in moderating content so as to protect users from harmful content and online abuse? Secondly, what measures should, and indeed could, be instituted by platforms to improve their content moderation and their complaints procedures?
Mark Bridge: If you are talking about sites like YouTube and Facebook, they are increasingly successful in using algorithms, which are quite crude, to identify some of the most egregious content. They are good at picking up some of the jihadi material and a lot of the paedophile stuff. I am talking about on the big, main social media platforms. What tends to happen is it gets pushed on to smaller platforms and smaller websites, to some extent. We wrote this weekend on a report by some analysts in the States, which showed that jihadi material was being pushed off YouTube but a lot of it was appearing on Google Drive and Google Photos, for example.
They have the AI technology to make some progress on this but they tend to rely on artificial intelligence, on the one hand, and human moderators looking at flagged content, on the other hand. That is content that has been flagged by users or, in the case of Google, trusted flaggers, which are organisations they work with. My sense is that the number of humans involved in this is still far too low given the enormous resources we have spoken about, and the degree to which they will proactively look for any of this stuff is, again, fairly small.
I have one quick point on that report I mentioned that we covered, showing stuff was being pushed from, say, YouTube on to these smaller platforms. In that report, they made the point that the resources of your Googles and Facebooks should be used to back up the efforts of the smaller companies that do not have the same kinds of resources. They have not only the financial resources but the technical ability to assist a lot of those smaller companies that need to be involved in this as well.
The Chairman: Mr Hern, is the failure to resource human moderating a key part of the problem?
Alex Hern: It is, yes. It is slightly unbelievable that any platform with users measured in the billions can count its human moderators in the thousands. That seems to be a scale error. It is not the case that you need someone to read every piece of content that goes up. I do not believe that is what anyone is requesting but, none the less, more people helps. More people allows, first, quicker turnaround times, which is the obvious request for issues like child abuse imagery and terrorist content. More people also allows longer, more nuanced, more considered review, which is of course a symmetrical problem. For everything that gets left up wrongly, there are things that get taken down wrongly as well.
I am sure all three of us have had situations where we have reported both of those things to the companies, only to find that miraculously when a journalist is spotted, they reverse their decision. One goal for moderation should be that happening less often. It is odd that journalists, without any of the in-house tools to mechanically search through these things—a journalist just uses the public entry points—can so consistently find examples where the policies have been implemented wrongly, either too strongly or not strongly enough.
Beyond simple resourcing, an almost larger issue is transparency. The problem is that Facebook has published vague versions of how they apply their rules. Most of the other companies have not even published those. It is still very unclear how any of this actually happens. Without that, it is hard from the outside to really recommend anything. We do not know. They talk about themselves as though they are operating miniature states with judges and juries but, if they are, it is sham justice.
Matt Reynolds: I completely agree with all that. To add from my own experience of reporting, I did a story around some really nasty far-right pages on Facebook that had a reach of millions of people in the UK. Obviously I spoke about this to Facebook. They were not taken down because they did not apparently violate the community standards, even though they were talking about siphoning funds to militia groups. It was very, very bizarre. A couple of months later I then found out that they had been taken down. I got back to Facebook. It was very hard to get an answer from Facebook as to why this happened, but the answer was that a post had violated their terms. We are talking about a network of 12 pages with millions of people.
Like Alex was saying, there is not this consistency. That is a big problem. There is not this transparency. Furthermore, is not applied uniformly when it comes to sites of that scale. Especially when it comes to sites like Facebook, you have to remember that groups’ reach and success on these networks are how they make their money, right? It is a problem. If these rules are not enforced uniformly and scale comes into it, you start to wonder, “Well, it should not be a factor who commits the crime or violates the hate speech”. In my experience, it seems to be that those are not uniformly applied. It makes my job quite difficult, to be honest. It worries me slightly.
Baroness Quin: I am struck with what Mr Hern was saying about them not publishing their own rules. Are there ways in which they could be obliged to publish their rules, either through the EU, if that is big enough to do it, or through consumer demand? Lots of petitions get organised for all kinds of things. Is it conceivable that one could have a petition of angry consumers saying that they wanted to know exactly what these rules were?
Alex Hern: That is what I was going to say. This takes it back to one potentially fruitful avenue of regulation. If a regulator’s job is not to sit on the outside and attempt to moderate these companies themselves, which would be horrendously expensive and very hard to do externally, a good avenue to explore would be a regulator whose job it is to ensure that these rules are applied consistently and fairly. Obviously one part of setting up such a regulator would be that the regulator would have to demand to know what the rules actually are.
It does feel that a regulator whose job it is to—in the German style—demand that hate speech be taken down in 24 hours or they will issue swingeing fines, is obviously going to result in an overcorrection; it is going to result in what has reportedly been happening in Germany so far, where anything that has the slightest whiff of hate speech is taken down. I believe that in the German public sphere that is not seen as that bad a thing. Here we have less of a hard-and-fast desire to fully censor all potential hate speech from society. We are a more open society in that regard.
If rather than that we have a rule that says, “Your rules about hate speech must be clear, open to your users, applied consistently and applied through a moderation process that is worth its salt”, then you start coming to it being actually quite fair to fine for breaches of that. If you find out that Facebook left up some hate speech because it was from a user that had a million subscribers and contributed quite a lot of video that generated ad revenue, that is the sort of thing that a regulator should be clamping down on. It is a fruitful avenue.
Q157 Viscount Colville of Culross: I would like to declare an interest as a series producer providing content for the Smithsonian Channel and also for CNN. I would like to ask you about fake news and your views on the significance of fake news in threatening the trust the public have in online media. What is to be done about it? Alex, you have just talked about transparency. Matt, in your Wired article on the fake news interim report from the House of Commons, you said that the report was an “indictment of technology companies’ opacity”. Should we get any future regulator to enforce transparency so we know where the sourcing is and we know where the users are? Can you tell me about that?
Matt Reynolds: I certainly think that in terms of the sourcing, yes. What fake news or misinformation has done does not just affect online media but more broadly it has played into this quite cynical distrust of authority or news more generally. I would say that at the very least it makes an awful lot of sense to flag up the source of where information is coming from. Although, to be honest, I wonder whether that would necessarily solve anything. If the source is Tommy Robinson and that source does not use facts or that source is erroneous in their reporting, does it matter if it was Tommy Robinson? In fact, it is the virtue of the brand of Tommy Robinson that attracts people and not their reporting credentials, if you like.
Sorry, I realise that I am not offering a solution, but I think that the problem is that when you say, “We need to factor in all that stuff”, I completely agree, as a value judgment for what people say news should be. The big problem with Facebook is there is news, not news and all these blurs. It is almost like you have to make a decision: “Is this thing presenting itself as news? Then it has to meet these standards. Then, should it be somewhere else?” I cannot see exactly what you would have.
It is a good idea to have more transparency, and there should be more vetting and more openness. It is a really complex problem that would not necessarily just be solved by saying where it comes from. We need to decide what information is useful and what context is useful to provide. Do we talk about who is funding it? Do we say that they have been brought up for this violation before? There is all this stuff. Yes, there are lots of questions before we get there.
Alex Hern: Fake news and misinformation online is an extremely difficult topic. It is hard to even define the terms of the debate. When fake news first started being used as a descriptive term it was referring to sites that published stories, created out of a whole cloth, to gain revenue from adverts that were run on them, and were shaped largely to talk about American politics because that was where the most readers and the most obvious readers for shock news at the time came from. That is, as a category, quite easy to deal with because, for one, the ad networks—networks such as Google, who effectively fund these sites—do not really want to fund these sites and did in practice begin pulling adverts from them quite rapidly. Again, in terms of what we were talking about earlier with older regulations kind of still working in the internet age, much of what these places publish is libellous and defamatory, and they do not hold up very long if anyone wanted to sue them. They also have a short shelf-life.
The broader problem is the lack of public trust that they have engendered. In short, the broader problem of fake news is that now people call real news fake news. I do not think there is an easy way to solve that. I may sound like a company man when I say it, but one way is for the masthead of the site to start mattering again; and for people, when they are judging whether or not they believe a story and whether or not they trust it, to partially base that on whether or not they trust whoever is reporting it. I like to think that the Guardian readers, by and large, trust the Guardian stories, in part because they are from the Guardian, because we have a several-hundred-year reputation for doing that and for providing trustworthy content.
What seems to be lacking is the opposite side of that, the moment of, “Hold on. I have never heard of whoever is reporting this before. Should that be a red mark? Should I hold my fire before repeating this to friends and family?” That may be a cultural thing that will pass. We are still quite early in this. We are very early in the majority of the British polity gaining a substantial proportion of their news from the internet and, beyond that, from social media. It may be that people will just learn and change their patterns. That is a very hopeful, optimistic view of things, but I would almost rather sit on the side of optimism than on the side of a lot of potential interventions like state bodies or the large platforms having to verify whether things are true or false or algorithms and machine learning systems trying to learn automatically to verify whether or not a thing is true or false.
If I, as a journalist, go out and interview someone and they say something new to me and I publish it, broadly there is no automatic way of verifying that. It is the nature of reporting that, to a certain extent, you are taking it on trust from me that I am accurately reporting what I was told. You cannot send an algorithm off to see if that has been reported anywhere else, because I was the first person. You cannot read those comments and then compare whether or not what the person was saying stacks up next to a Wikipedia article that was on the same topic, because the person said it and that is the news story. Reporting has come down to trust for a long time. Sites, companies, journalists and individuals who lie tend to get caught out and that has a damaging effect on their trust going forward. I hope that that will continue to be the case in the online world.
Viscount Colville of Culross: I do admire your optimism very much indeed. It does not seem to be going that way. It seems that there is more and more fake news taking place. I have been a reporter, and when you interview someone you do make sure that what they say has some fact and some truth to it. That is the point of the media: you mediate it and decide whether or not there is some basis.
Alex Hern: On that point, if I may clarify my remarks, I mean that the very fact that person said something has no external source of fact-checking. If someone accuses me of making up a quote from you because they cannot find that quote anywhere else, that is not mechanically checkable.
Viscount Colville of Culross: In a world in which we are seeing fake news that seems to be growing exponentially, contrary to your optimistic view, is it not about time that we had some sort of intervention and some sort of determination that there should be some fact-checking that, “Yes, this should be allowed to go viral”?
Alex Hern: That would be hard. That is effectively the issue here. We say “go viral”; we are talking about social media mechanics. Fake news can come from any individual user on any one of a dozen or so social networks. It is not clear to me what checking there would be if we are to continue to have social networks.
Matt Reynolds: My fear is that a lot of the popularity of things that do go viral or the people who say these things is precisely because they are not part of those organisations and they are outside of that bubble and that type of thing. Although I completely agree with Alex that the idea of authority in a masthead is something I believe in and I would hope, broadly, people come around to again, I think that a lot of what we are seeing is a very direct reaction to that. Trying to put another authoritative wraparound on that would not necessarily solve the problem because I do not always think that people are critiquing it in the same way. I do not always think it is a lack of information. This is a problem with the wider public debate: that it is about being acerbic, it is about being contrary and it is about being different, and actually contextualising it with facts might just not help. It might just seem like another authority that you can ignore. Sorry, I realise that is not a solution. That is the pessimistic side, I guess.
Q158 Baroness Kidron: I really wanted to ask you to think about it in a slightly different way, in terms of whether you have any attitude towards the design of service. For example, a YouTube insider recently said that 70% of YouTube videos are shown on the “recommend” button. That means that YouTube has an immense power in the direction of travel of what people are seeing. We also know that those things get more and more extreme. If you start as a teenage girl on a diet site, you end up on a pro-ana site and everything else. We appear to be discussing it as if there is Person A saying one thing and Person B saying something else that is true and they have an equal chance. I really would love you to all say something about the design of service and what responsibilities lie in there, because that does seem a more fruitful place for regulation.
Alex Hern: To be clear, it is that last aspect—whether or not it is clear for regulation is my problem. There are a lot of things that services can do to help with this. Facebook, for instance, has made strides in this direction but for a long time, and particularly in 2016, at the height of or birth of this fear of fake news and misinformation, Facebook radically deprioritised information about where the news was coming from, about what site the link was going to and emphasised information like which of your friends had shared the story and what they had said about it. That strips all of the anchors of trust that I was talking about earlier: the fact that it comes from a credible news organisation, the fact that the original headline was written by a professional journalist, and instead it replaces those with its own sources of trust: this comes from one of your friends and they, until recently, had the ability to rewrite the headline on the link, which is a terrible idea for a news service.
All of that does not help. Facebook has made strides in this direction. They have improved that but they still prioritise the person who shared it because they are a site that is increasingly about fostering connections between friends and family, rather than connections between the corporate world and individuals.
Baroness Kidron: When they are fostering friends and family in a way that creates ethnic cleansing in Myanmar, is there not a responsibility within the design of service about quality of information, et cetera?
Alex Hern: There is, but if there is a more obvious criticism about Facebook in Myanmar, it is that Facebook does not have any staff in Myanmar. Facebook launched in Myanmar without bringing moderators to Myanmar. That feels like a very obvious first step before we start talking about legislating design of the service.
Mark Bridge: There is a responsibility. This is based on algorithms. All of this comes down to their algorithms, which again are a black box. We do not know how they work. There needs to be more transparency on how they work.
On fake news and misinformation, brands like Wired, the Guardian and the Times do retain a lot of trust. However, if you are a smaller player coming out now, and you are decent and doing the right thing, you are going to suffer because people do not have that confidence in you, because of the fake news. The smaller, less known journalism outlets are the ones that are going to suffer more.
Education is really important in this. We have a programme going into schools showing them how journalism is done, how facts are checked and those kinds of things. Again, you have to prepare people, so there needs to be work in schools to get kids familiar with this stuff.
The Chairman: We need to move on to another equally fascinating subject area, which is market concentration.
Q159 Lord Gordon of Strathblane: In a way, for all that one might deplore the way internet companies gobble up potential competitors, it does seem almost inevitable that once you have acquired a 51-49 lead in any segment, it very rapidly becomes 90-10, or almost total dominance. That total dominance sometimes can act in the public interest as well. Can you envisage Google being split? A to L, you use Google; N to Z, you use something else. It is kind of unthinkable, is it not?
Alex Hern: I can easily imagine a Google being split so there is a search engine, an email service and a video hosting site.
Lord Gordon of Strathblane: Take that slowly.
Alex Hern: A search engine, an email provider and a video hosting site. If one were to sit down and decide to split up these companies, first, one would have to be either the American Government or the European Union, because there are very few other people who could enforce such a change; secondly, there are clear lines of cleavages. These are large conglomerates. They are not an individual service. Facebook would split very easily into Facebook, Instagram and WhatsApp. Google would split very easily.
Lord Gordon of Strathblane: If we concentrate on Google just for a moment, might it not then lead to you having to pay for any search? Search itself cannot be monetised. They get their revenue from the Gmail and other things.
Alex Hern: Search itself is actually extraordinarily profitable for Google. In terms of search adverts on some topics, a single click can sell for $70 or $80. A search indicates a desire on the part of the customer to find information about that sort of thing. Most obviously, if you search for something like how to buy a car, you are a very obvious target for advertising about how to buy a car, far more obvious than almost anywhere else in the pre-internet age. Search is extremely monetisable. There are areas of Google that probably are not, and that are cross-subsidised by it. One way of looking at that is that is the benefit of running a monopoly: you have a lot of money to cross-subsidise extension into other areas. I am not wholly convinced that that is, in the long term, good for a competitive marketplace and broader innovation in the technical sector.
Mark Bridge: I agree with that. At the moment, as I think Matt said earlier, I use Facebook and WhatsApp—most of us use Google services and Facebook absolutely all the time. All the harms we have talked about linked to these companies are magnified by this ubiquity. The volumes of data that they can collect through this, and the hold they have on every aspect of our lives, again raises all sorts of concerns.
Matt Reynolds: If you are looking at the situation of Google, where Google has leveraged its position to squeeze out competitors on Google Shopping, it was fined for that; it continued to do it up until now. While I agree that scale in the abstract sense is good for me because I can search, do not pay for it and that type of thing—so I agree it has been useful in the same way as the internet is useful—I do not trust those companies not to leverage their power to squeeze out competition in other areas, which is ultimately to the detriment of the users. Yes, I am pretty cautious on that.
Lord Gordon of Strathblane: Mr Hern, you said that Google could be split up. Do you think there is a moral difference between Google using information on my taste from what I search for on the search engine and them scanning my Gmails to see what I have said I happened to like to friends so that they can advertise to me?
Alex Hern: One could argue that there is in that the search advertising is more based on an express desire to see a certain type of information, which ideally search adverts are then presented against, whereas email scanning to present adverts is much more passive. I should note that Google no longer scans emails to present adverts. It did for a long time. They changed their policy on that, broadly, it appears from the outside, because they were not making that much money and it was very bad in PR terms to mechanistically read people’s emails, so they ducked out of that market.
The Chairman: Baroness Chisholm, you might have a partial solution.
Q160 Baroness Chisholm of Owlpen: You all have talked about the clustering of power and centralisation. We have heard before in this Committee about the dominant platforms and how they benefit from their extensive data silos. I wanted to ask you whether more data portability could help control the power that dominant platforms exercise over personal data.
Matt Reynolds: I agree that data portability should be a fundamental right and you should have the ability to do that. It would very useful if you could say, “I want to download all my photos”, which you can now do under GDPR, or, “I want all the information they have about me”. I am not that optimistic that data portability would enable the creation of a new Facebook. If we are looking at the scale that Facebook is—2.2 billion people—why are you on Facebook? You are on Facebook because all your friends are on Facebook. You are not on Facebook because you can take that away. Maybe if I could take all of my friends with me and I know they would get that, perhaps it would be a positive thing. Fundamentally, a lot of these things are useful because of the scale, and they leverage that scale more than they leverage the cumulative data to keep people in. While I agree with it in principle, I do not necessarily feel that confident that it would actually change much.
Alex Hern: It is certainly true that the lack of portability has made it harder to build a competitor. The most famous example of portability helping is that Instagram grew as an adjunct to Twitter. For a lot of Instagram’s early days, you would sign up on Instagram, it would scan your Twitter followers and see if there was anyone on Twitter who was also on Instagram and you would follow. That was great for building a social network. I am not convinced that in the internet of 2018 the largest social networks would be easy to compete with if only for data portability.
Instead, we should sit down and say that in the market that Facebook is in, it is quite hard to conceive of a competitor coming at it head on and taking it on. Rather than trying to enable a Facebook competitor to build up and use data portability to create competition, maybe we should instead say, “Fine, Facebook has a natural monopoly; the network effects mean that it is extraordinarily valuable and any competitor would really struggle to deal with it. What does that entail for regulation? How does Facebook being a natural monopoly change the Government’s responsibilities, or the responsibilities of Governments in general, in shepherding it and protecting its users?”
Mark Bridge: In principle, it is a good idea. Again, I am not optimistic it will be that helpful. Sir Tim Berners-Lee is creating a new platform at the moment where your data is stored in a kind of pod and then different services and different apps get access to different bits but you control that; you see what data is there in your pod and you let an app have a certain amount of that data. It will be very interesting to see how that works out. If that works, that is fantastic. Again, the big incumbents, Facebook and Google, are so convenient. That is the thing: people will trade a lot for convenience.
Baroness Chisholm of Owlpen: Matt, you talked earlier on about how the big platforms are reaping the benefits of their scale. Do you think that should also bring some form of responsibility? Should they step up to the mark, basically?
Matt Reynolds: Yes, undoubtedly. It has to be that. There should be a very high expectation that once we have decided what they should do, they should be held to it and face the penalties if they do not meet that.
I would add to Alex’s point. He talked about natural monopolies and Facebook not having a head-on competitor, which I completely agree with. We should also be thinking about what the next platform is. What about in the home? At the minute, look at the companies that are creating the devices for voice communication in the home: Google, Amazon, Facebook and Apple. The reason is because you can leverage your dominance in one market to then get the second one. Rather than thinking about how you create a competitor to Facebook, perhaps we should also be thinking about how we make sure that the next platform is opened up. Maybe that is not about personal data portability but it is about training data and enabling people to develop these things. At the minute, Google is thinking about protecting search but it has search. It is thinking about what people are going to be doing in 20 years and how it can capture that. We should be thinking about how we keep that field open as well.
Alex Hern: On that note, with hindsight it is very easy to say that the single greatest failure of tech regulation in the past decade was allowing Facebook to acquire Instagram. Instagram was probably the greatest risk to Facebook’s monopoly that the internet has seen to date. It was succeeding in precisely the way we have talked about: it was not coming at Facebook head-on. It was slicing off a part of Facebook that people engage with very strongly, which was the photo-sharing part, and creating a social network that could quite healthily run parallel to Facebook. Facebook bought it for $1 billion, and then bought WhatsApp later on for $13 billion, and two potential avenues of quite fierce competition with Facebook were cut off. Hindsight is wonderful.
The Chairman: We need to move on. We always finish with a little bit of Brexit—it is the nature of this place; we have not got away with it. I am going to ask Baroness Quin to read a question to put on the record and then I will ask our witnesses, if they will, to reply in writing to Baroness Quin and the Committee.
Baroness Quin: As the Chairman says, it is impossible to be in the Palace of Westminster for a meeting and not hear the word “Brexit” uttered. What effect do you think the UK leaving the European Union will have on the regulation of the internet? Will the UK lose influence as a result of not having a seat at the table? What do you think the overall consequences of that would be? If there are negative effects, how can they be mitigated in the future?
The Chairman: The clerk will confirm that question to you and, if you would be good enough, we would ask you to just send us your thoughts on that. Your evidence has been very helpful to us. You have answered some questions and raised some other ones, and you have explored some areas that are very interesting and useful to the Committee. When you write to us, if there is anything that you feel that we might have considered or that you would have liked to have elaborated on, your thoughts would be very welcome and will form part of the evidence before the Committee. Mr Bridge, Mr Hern and Mr Reynolds, I thank you very much for giving evidence to the Committee today.
