Examination of witnesses
Robert Colvile, Jamie Bartlett and Laurie Laybourn-Langton.
Q52 The Chairman: Can I welcome our second set of witnesses to our evidence session this afternoon on our inquiry into regulation of the internet? Our witnesses are from three prominent think tanks who are working in this area: the Centre for Policy Studies, the IPPR and Demos. I thank the three witnesses for being here. The session will be broadcast online and a transcript taken. Can I ask the witnesses to briefly introduce themselves, tell us a bit about their organisations and perhaps start by giving their initial impressions on the issue of regulation from an economic perspective, the dangers of poor regulation, the impact that it can have on start-ups and innovation in the sector and the likelihood of big tech companies locating in the UK if we create the wrong kind of regulatory environment?
Robert Colvile: I am the director of the Centre for Policy Studies. Before that I was a journalist at the Telegraph for 10 years where one of my main areas was technology. I wrote a book called The Great Acceleration, about how the internet is speeding up the pace of life, including politics and the media. I then migrated into think tank-ery. Our organisation is a free-market think tank devoted to policies that promote opportunity, enterprise, aspiration and ownership—and, obviously, the internet is now threaded through the economy completely. In terms of regulation, the starting point for Britain should be that it is in a good place. It dominates Europe in terms of tech investment. We have a higher percentage of tech jobs in the UK than most other countries.
There are issues around the fact that the UK cannot grow its own Facebook and Google. I have written about this in the Financial Times and elsewhere. If the starting point, as with a doctor, is to do no harm, that is something that definitely needs to be borne in mind. There are tremendous issues and challenges thrown up by the internet but one of the dangers is regulating for the internet as a thing rather than thinking about it as a whole host of areas and activities which are deeply entwined between online and off—similarly, regulating for Facebook or Google and accidentally catching up with the rest of the economy in the process.
One of the things that I want the Centre for Policy Studies to get into under my directorship is the issue of monopoly policy because everyone recognises that we are in a situation now where the traditional template does not apply. The idea is if Airbnb were to get 80% or 90% of the market, but was still driving down prices for overnight stays, is there consumer harm and what can you do about it? What is the benefit of the traditional template of that? It does not work. I believe that is a fascinating area that we will have to explore over the next few years.
Jamie Bartlett: About five years ago I set up the Centre for the Analysis of Social Media—CASM—at Demos with the idea of trying to take techniques of machine learning that were being applied in the private sector and work out how to use it in academia and public policy research. We partnered with machine learning specialists at the University of Sussex and developed software, and methodologies and techniques for using those technologies in research work. I am the author of a book that is rather scarily titled, The People Vs Tech: How the Internet is Killing Democracy (and how We Save It), though the “how we save it” is in brackets, so it is almost an afterthought, unfortunately. I am especially interested in the ways in which digital technology is, in some senses, incompatible with modern representative democracies and how economic change driven by the internet will wear away at the fabric of the middle class, for example.
In respect of the question, even though my book would suggest that we do need to regulate more, I am especially worried about the risk of bad regulation here because the internet will change rather a lot in the next decade or so. It is changing very quickly already. There is a great emergence of censorship-resistant technologies which will make the way we understand who is responsible for content potentially change quite dramatically, and it will be very easy to pass very bad laws about how the internet works now, not thinking about how it might work in future. One of the great risks I see driving this is the great deal of political consternation about the role the internet is playing in politics at the moment. Some of that is driven by the way traditional news outlets are frustrated by their loss of advertising revenue to the big tech firms, which is causing some remarkable headlines that are not particularly helpful.
The final point about this is that I consider we are in a race with countries such as China over artificial intelligence and it is quite important that we win that race. China does not worry so much about user privacy and data protection and is investing a fortune in this. We have to find a way of regulating various aspects of the internet, as Rob said, not all the same, such that people still feel that it is democratically accountable in some way and they have some control over it, but that we stay ahead of countries such as China in the race for AI, and that is a very, very difficult thing to do.
Laurie Laybourn-Langton: I am a senior research fellow at the Institute for Public Policy Research where I work on the IPPR’s commission on economic justice, our flagship programme, which is creating an economic platform for post-Brexit Britain, of which the internet and, by extension, platform companies are key parts. I have worked in economic policy for around five years, prior to which I worked in digital campaigning looking at how to translate digital tools for social action. Our main focus at IPPR when it comes to regulation of digital and the internet is platforms—those intermediaries of social and economic activity.
Our main point here is they have disrupted socioeconomic relations and the way we do things in the economy and society more than many other inventions in recent history and, arguably, as time goes on, more than at any point in the last few hundred years. Therefore, an appropriate response to that disruption is key and, we believe, ranks up there with the major challenges of our time, including the environmental challenge, pervasive inequality and many of the others that we often hear about.
On the regulatory challenge, the first point I would make is that we still need to work out what we are regulating, because of the extent to which these disruptions have worked into a variety of economic and social activities and some political activities as well. By its very nature, regulating becomes very hard because of the fast pace at which this disruption is occurring and permeating into markets that we do not necessarily associate with the first movers within this sector.
There will be an enormous cost in getting regulation wrong and there could be an enormous cost in getting it right, which goes to the heart of this problem around the power relations between the interests that are served by a lot of the first movers, the large companies having large interests increasingly in a number of sectors. As with many areas of regulation, we will probably not know whether and what kind of cost we have incurred until far into the future, which becomes particularly dangerous for a number of reasons, including around AI, which will increasingly undergird the infrastructure in which economic and social activity can occur in the future.
Q53 Baroness Kidron: One of the things that concerns us is that the conversation easily floats to content when the infrastructure itself, the design of services, the ways in which they work are the things that maybe we should be looking at more carefully. Could each of you say a bit about your thoughts or worries about some of the current norms of design, by which I mean filter bubbles, persuasive design technologies or the idea of machine learning, where the designer does not know what the outcome will be? Going to Mr Colvile’s point of “do no harm”, should we be designing them to do no harm and considering in advance what they may do? Could you speak to the design?
Robert Colvile: These sites are designed to succeed and, primarily and overwhelmingly, to foster user engagement, revenue, whatever the goal of the company is. They are ferociously well-equipped to do that by a process of constant iteration. David Cameron does march into Tesco and tell people not to put their chocolate oranges by the checkout, but he does not start telling them to shift the vegetable aisle 20 metres to the right and all sorts of other things. We need to be wary about doing things to interfere with the activities of online companies which we would not do offline because they happen to be on this thing called the internet.
Baroness Kidron: May I press you a bit and perhaps talk about the drug sector, big pharma, which does have bars that it must reach. No one tells them what to put with what, but they do have to think about the consequences of their design. I did not mean literally tell them how to design—but perhaps impact and consequences would have been a better way of putting it.
The Lord Bishop of Chelmsford: I believe sweets by the shopping tills is an extremely good example and not a trivial one at all, because I used it on a previous occasion. As a society, that is where we legitimately make decisions about design because we think there is a societal good to that.
Robert Colvile: The problem with any of these debates is that you get into a whole mess of areas. There is a conversation to be had about Facebook as a mechanism of constant engagement. The example I have used before is that if you “like” UKIP, which is a perfectly legitimate political party to like—people in this room might dislike it—it instantly shows you pages for the National Front and the BNP. It is constantly trying to intensify and radicalise your experience at the same time. I think and hope that they fixed that in the year since I found that, but it is a good example. You have questions of structure with things like Amazon—if you own the platform, should you be allowed to be a player on the platform? Ultimately, these companies succeed or fail by how well they serve their users; that is the metric. They are far more scared of doing stuff that would disappoint their users or lose their loyalty. That is the thing which drives them on. It is not a sinister plot to hypnotise our children into using Facebook. It is just that they are paranoid that they need to keep people happy and engaged.
The Chairman: It is about how good the markets audit.
Jamie Bartlett: Possibly. I am slightly more worried than Rob about addictive technology and design technology, but I agree that it is not a malicious plan. It is just that there is an impulse to work out how to make people spend more time on the site to collect more data. It is the underlying driving logic of how these platforms make their money. However, I do not believe that you should be in the business, therefore, of telling companies how to design platforms.
In this instance, I am talking of a self-regulatory, ethical body where companies sign up saying, “We are designing these platforms as something like a fair trade stamp whereby we are not designing them to be as addictive as possible. We are going to try to put in some default settings to make sure you are reminded that maybe you have been on this platform for an hour now and do you want a break?” There are some tech firms that do those things, such as Slax. Default settings is quite an interesting area where the default setting of whether your data is immediately shared or not has probably more influence and impact than any other small tweaking around the edges. There is scope for it to say, “Well, could you have your default setting at higher levels of privacy rather than at lower levels?” There are certain things that can be done.
It is interesting to talk about it now because advertisers talk about the “creep factor” in their work—an invisible line whereby it seems a bit weird and people do not like it very much, but understand it to exist. At the moment, it feels like we are on that line with addictive technology and a lot of tech firms have been quite explicit—including Mark Zuckerberg—that they feel they may have crossed that line and want to work on improving it. At this point, it is a good opportunity to work with the tech firms to try to do something about it.
You mentioned filter bubbles. It is an idea whereby like-minded people are clustered into silos of information and it radicalises them in one way or another, which is not entirely my experience, looking at the research base. We do have access to a lot of different views. When I am on the internet, sometimes I feel that I only ever see different views and they are all wrong and that is the problem. It is not about thinking that you are stuck in a filter bubble of like-minded people. There is a broader issue. Simply the way we communicate with each other online is very sharp, quick, and dramatic. We tend to overstate our enemies’ or opponents’ importance and significance, and we attribute to them all sorts of terrible motives that they probably do not have, and they do likewise to us. To me, that is a bigger problem than a filter bubble. Some people call that the backfire effect, which is the nature of internet communication. Essentially, it leads you to see a lot of other views but to disagree even more when you do see them.
Laurie Laybourn-Langton: The point I would make is that social, political and legal norms have been established over time in a number of areas of activity across the economy and society. Many of those standards are stronger for important types of activity, whether that be how we interact in democratic elections or in the provision of certain key economic goods, of which drugs are probably one—pharmaceuticals. I do not believe we have yet established those norms which are translated into regulation and policy for platforms and digital firms. A perfect example of this would be that today, the Times reported a story about how a video was leaked showing the development technologies that could lead to a better understanding of how we behave, to then pre-empt how they can manipulate our behaviour into the future.
If the Cabinet Office had a video like that presented to them and leaked, can we imagine the reaction that would rightly come about the potential for cracking down on people’s liberties across the country and, indeed, the world? That shows a big disconjunction between the norms we have established in certain areas of society and those in this sector, particularly when we understand the activities that we take part in on Facebook and other platforms are increasingly akin to key essential services for the economy of gaining information, conversing with people, building political coalitions and those in the community. In establishing those norms and the regulations and policies that are attached to them, I completely agree with Jamie that the revenue model that sits behind a lot of these platforms is the key thing.
They create certain services, which are underpinned by algorithms that sift through data; they extract data and analyse that to gain insights, which create products which they can then sell. In the case of Facebook that is advertising. The impulse there is to extract more data to feed back to improve the analytics to further extract data and provide products that gain profit. There are three areas which we are looking at around policy in response to that. There are particular ones about how platforms develop the algorithms and the services that underpin their model. Do we need to introduce certain types of ethical behaviour when it comes to those that designed coding and algorithms, because we know that, famously, you often translate your own political or other biases into the programming that you take part in? Do we need professional qualifications that have similar ethical support, as well as standards that chartered accountancy has in this country and others?
Secondly, we then need to make sure that certain norms are translated into regulations. Another article on the front page of the Times today was about how some auto-response Google searches are beginning to potentially link those who are implicated in sexual assault cases with the victims because people have leaked those details on social media. Google is able to take some of those things down. There is a norm that we have in wider society about sharing that information—a legal one as well—that increasingly needs to be transferred to regulation.
A third area is increasing competition in some of these markets. I hope we will talk about that more. The one thought I would add at this point is that, in doing so, we need to break down this whole world of the internet and digital technology into certain functions instead of it lumping together.
Baroness Kidron: I was going to ask you about whether you should regulate but you have answered that. All the witnesses, of all kinds, who have come here all agree there is ethical component. What we are struggling to understand is where that sits. If you could very briefly say, in the conversation about “We should have ethics and norms”, and so on, who is “we”?
The Chairman: This is something we want to get to. We have been asking witnesses about the “who.” There is clarity about what the industry needs but not about whether there is regulation, law and who is responsible for bringing this together.
Baroness Kidron: Because they suggest they would like to be the author of the ethics. That is complicated in itself. Where does that duty sit, the creation of an ethics context? Anybody? Any brave man?
Jamie Bartlett: As I have said, when it comes to addictive tech in particular, it is a question of self-regulation. It should be encouraged and welcomed by government, but, in the end, we should at least first try with self-regulation.
The Chairman: Can you help us with the steps to encouraging the industry? You talk to the industry and have an understanding of where they are coming from on this and, from your answer, you are optimistic that they understand it and want to get to the same place. What steps should be taken by government to make that happen with the industry?
Jamie Bartlett: I do not feel well enough qualified to answer that. You are probably more experienced with the ways in which governments have been able to encourage good self-regulation. In the alcohol industry, they bring a large body of players together, talk about the different ways in which they probably should try to self-regulate better and if they want to set up a body, the government will be there to support it and there is the threat of regulation if they do not do it. That seems to be the way that it usually works.
Laurie Laybourn-Langton: I will make three points on this question of who is one of the most important ones. First, this is enormously complex. We are talking about platforms that engage in activities that span from payment services all the way through to providing a way to have a political argument, build a coalition, take a petition to Parliament. The regulatory response is not going to catch all that in one go. We should break it down into the particular activities under the purview of these services.
Secondly, within that, there needs to be future focus. Your regulatory response to whether you can, without impunity, place what we would probably define as hate speech in a public forum out there—the regulation or any approach we have will be vastly different from potential regulation to stop Google, for example, from developing technologies that are able to nudge or manipulate us in a way that we have not yet fully understood.
Thirdly, a kind of direct answer to the question of “who”, I would say, is: not a small subsection of the leadership of the firms who provide these services, necessarily. A more general version of that is that, among the key regulatory challenges, this should be undertaken according to democratic principles, in the same way that we have provided regulation in other key areas of society and the economy through a democratic mechanism—Parliament and the people who represent us—as opposed to those whose revenue model dictates that they should focus narrowly on products they think could revolutionise the world.
Lord Allen of Kensington: Would it be helpful or not to fund an independent body or to fund what could be done in this area?
Laurie Laybourn-Langton: Taxation has to be part of the toolkit, but, again, it depends upon which particular activity you are seeking to regulate. For example, you could tax more heavily the provider of an on-demand taxi hire service in a city if it does not list vehicles that are electric.
Robert Colvile: It is generally accepted that, at least in terms of the early years, the tech firms did not take responsibility for what was happening on the platforms. Mark Zuckerberg stood up again and again and said, “We are a platform, not a publisher”, which coincidentally and helpfully means they do not have to spend lots of money hiring people to police content. No one can police the internet, but it is probably fair to say they were not trying very hard. That has changed now as the backlash has grown. They are investing far more in that kind of thing. The Conservative manifesto in 2017 was littered with commitments to impose better behaviour, to push the internet companies to behave properly and make sure that what was illegal offline was illegal online, and to prevent hate speech. I have a printout here. It was quite a lot of stuff. In many ways that manifesto has been inevitable.
I do worry that there may be some slight rule of unintended consequences to some of this. We have seen it with the GDPR, which I imagine we may get on to later. The big companies can afford the lawyers and the squadrons of people policing all the comments and writing the algorithms to hunt down Islamist videos and all the rest of it. As with quite a lot of regulation in quite a lot of sectors, in many ways what that does is deepen the moat around them and make it harder for people to compete with them. In certain cases—Jamie has written about this—it may drive the more unsavoury behaviours into a wild west, which may be impossible for the authorities to see, let alone regulate, especially with advances in crypto-technologies.
Baroness Quin: What are your feelings about whether there is a need to strengthen the consumer voice in this process and, if so, how to do something about it? What are your feelings about the role of education? It seems to me that a lot of us—I certainly include myself—use the internet for convenience but, in accepting cookies and goodness knows what, do not understand a lot of the time how much information about ourselves is being shared in a way that perhaps, as individuals, makes us feel uncomfortable. I certainly feel in that position myself now, having done so much online and suddenly realising that in accepting all these cookies and so on, all kinds of information is out there. Related to that, Jamie mentioned having some kind of default setting that gave a higher privacy standard. Could that be done by self-regulation, or ought it to be legislated for? Is it covered in existing legislation? I do not know.
The Chairman: Shall we start with education and citizen empowerment, Mr Bartlett?
Mr Bartlett: I will clarify something. When I was talking about self-regulation, I was referring specifically to the point of addictive tech and design. As you said, Laurie, it depends what problem you are trying to deal with. On user education, when I first set up CASM five years ago, we wrote a big report talking about the value of and need for better education on digital media literacy in particular, i.e., the ability to distinguish between truth and falsehood online. Every single report that any think tank ever writes ends up saying, “We need more education”. The curriculum will be 150 hours long a week to include all the content we need.
Of course, I think that but I accept that will never be enough. It is an easy one to default to. There are certainly ways in which we can improve the way media literacy is taught. Google’s algorithms are not as powerful as our own cognitive biases in terms of how we filter information. If we have media literacy classes improved, it needs to include cognitive psychology rather than just technology. There are certainly things we can do but it will never be enough.
Interestingly, I do not think any company has done more for the cause of user privacy than Cambridge Analytica. Everyone is now talking about it and thinking about it in a way they were not before. The way that the population tends to learn about these things is through things going badly wrong and us taking the opportunity to think about it and have a bigger public debate about it. There is a certain way in which public education moves independently of schools. That is very healthy and has been quite a good thing over the last couple of weeks. Honestly, when it comes to legislation on default settings, it is something I am not sure about. To be honest, I cannot decide. GDPR has some things on that, but it is a difficult one.
Mr Colvile: On education, I would echo what Jamie says. Obviously, we need to increase digital literacy, but that is much harder than writing reports saying we need to increase digital literacy. There is some encouraging evidence that young people are more privacy savvy than their elders, and are using Facebook in ways which preserve their privacy and migrating to other networks because they do not want their parents finding out, which is perhaps the most important form of privacy when you are that age. Regarding consumer power, consumers have voted in overwhelming numbers. Some 70% of the UK now uses Facebook.
Mr Bartlett: A high proportion of people also say they do not trust the company or they do not like the company.
Mr Colvile: I agree with that, but if you look at the results that came out after it had this amazing storm of bad publicity, guess what? No one is giving it up. Consumers like these services, which is part of the problem because it is hard to regulate something when everyone is using it and is quite happy with it. The person who stands for election on a manifesto saying, “We will break up Google’s monopoly”, will be horrendously unpopular.
The privacy by default issue is interesting but there is a tension here between privacy and progress. I am not talking about Facebook in particular. Things such as NHS data in particular—doing the cool stuff with AI requires access to very large data pools. As Laurie has pointed out, we may not want them doing some of the creepy cool stuff. The value to Google is not that I am searching for a Chinese meal at this current moment or even that I tend to like Chinese meals; it is that people of my age, weight and location tend to like Chinese food and in 20 years’ time we will all die of heart attacks. That is the real goal that you are trying to reach in quite a lot of this.
The more privacy rules you erect, the more—this is not saying we should not give people as strong privacy as they want, but we need to recognise that the model relies on data being shared and analysed and that does good things for the rest of us. If you said to people, “You can have higher privacy settings on Facebook but you have to pay £10 a month to use it”, that is a very different proposition from, “You get the privacy, but they don’t get the data”.
The Chairman: I am sorry. We need to move on. Mr Laybourn-Langton.
Mr Laybourn-Langton: Education: there is lots of low-hanging fruit there, particularly regarding those who did not necessarily grow up with these technologies. The consumer point that Robert made is important and goes to the heart of the regulatory problem here. The original rules of regulation or ways of looking at the world are around price and impacts upon consumers. At the first instance, this is all very pleasing for consumers and this is where you probably have limits to education.
Can I make another comparison between a service and people’s ideas of liberty and government? If, one year ago—five, 10, 20, 50 years ago—you suggested that you would voluntarily put a device on your person that mapped where you went all the time and you openly allowed it to infer, based upon that data, where you lived, where you shopped, where you worked, where you did all sorts of other activities, the uproar would have been extraordinary. Because we have a lot of pleasing elements for consumers, it means that that disjoint exists. That is a serious issue that we all have to deal with. That is not about regulation; it is about how we interpret society.
That regulatory response needs to be properly resourced so that experimentation can occur. It can be asked whether we should have default settings and the like. We should experiment; that is the very nature of this sector. It is very fast-moving, it is experimenting all the time, and there is potentially room for regulators in certain small areas to experiment around certain products even in a safe atmosphere. I know this is increasingly happening in the financial services industry where regulators are looking at how to experiment in what they call sandpits, where they can look at certain products and what that means.
Q54 Lord Allen of Kensington: To what extent should online platforms be liable for the content that they host? Specifically, should the safe harbour provisions of the e-commerce directive be removed?
The Chairman: You have already addressed some aspects of this, so brief answers would be fine.
Mr Bartlett: No, they should not. As with publishers, it would more or less destroy most of the social media platforms if they were liable for all the content that was hosted there. The task is to make them as quick as possible at removing content that is identified as illegal, twinned with some self-regulation. The German model is a reasonably good one: high fines where they do not remove content within 24 hours if they have been told there is illegal content there. They face fines of up to €50 million. There is backstop legislation that encourages them to be much quicker and more responsible in removing content when it is alerted to them.
Advertising might be slightly different. Given that they are paid to put that content up there—this is obviously the Martin Lewis case that is going on at the moment—there is a case that that might be slightly different and that they should be liable for the content that they advertise on. One of the reasons why it is quite dangerous to go down the line that they should be liable for all of it and we remove the e-commerce directive safe harbour provision is that the internet is changing. In the next 10 years there will be a lot more decentralised blockchain-based platforms where it will be very hard for the companies that run those platforms to be able to remove content at all because they will be hosted on distributed immutable ledgers, at which point there is no technological capability that would allow them to even take it down. Therefore, for them to be responsible for that would be an extremely difficult case to make.
The Chairman: Mr Laybourn-Langton, I am particularly interested to know, do you recognise that distinction between content and advertising content?
Mr Laybourn-Langton: Yes. An equivalent would be that you could stand outside Parliament here and you could say something in that forum, or imagine a situation where Parliament Square would earn money from your ability to say those things. Those are clearly distinct areas, so I would agree with that. What Jamie is saying is absolutely right, and with new blockchain-based technologies it will become harder and harder to do this as time goes on. If you were to set that principle now, pressure would need to be applied to ensure they were moderating that content as quickly as possible.
On the question of safe harbour, we do not have a particular view on that but note that the gap between the value that YouTube is getting out of those who have created certain content and the money they receive in return is obviously opening widely. We do not have a specific view on that.
Mr Colvile: This will be shocking: the director of the Centre for Policy Studies will praise a European regulation. The safe harbour provisions work pretty well. There is an obligation to act expeditiously to remove or disable access to the information. I agree with Jamie that removing it would be disastrous. It is the internet equivalent of a limited company. It is what gives you the protection to start up a company and not be liable. Having edited a newspaper website, I know that it would have been apocalyptic if we had been liable for everything people said in the comments section. There is also an economic point: if you want a single way to guarantee that no one will ever create a new technology business in this country, removing the safe harbour provision would probably be one of the top three things you could do to make this a hostile place for people to invest.
Lord Allen of Kensington: Blockchain or whatever comes after it is in itself pretty scary, because we are trying to legislate and regulate for something that has not even been thought of. Do you have any comments or thoughts on how we can try to move ahead? Frankly, legislation is always too slow in these areas and your comments on that struck home to me.
Mr Bartlett: I am struggling greatly with this. In my first book, The Dark Net, I looked quite a lot at blockchain technologies. That was in 2014 and I could not see an easy way to deal with this. I do not think there is one. There is some enabling legislation or regulations that could be passed about how initial coin offerings should be managed and run. I know a lot of blockchain companies want guidance from government about how they should deal with some of these problems and there is none out there at the moment. For me, the first step is to create some kind of enabling environment for these companies to operate. Most want to work within the law; they do not know what the law is because there is not any law for it.
Honestly, when it comes to the issue of censorship in particular I do not think we have an answer. It is genuinely possible that we are entering into a world in which censorship of the internet becomes close to impossible. Governments will inevitably pass draconian measures to punish the people who posted that content because deterrence will become the most effective means of making sure bad material is not posted. That is the direction of travel I imagine it will go down, which worries me.
Q55 Lord Gordon of Strathblane: Moving on to competition law and its adequacy for the current environment, I was taken, Mr Colvile, by a remark you made in your opening comments about where the public interest lies if Airbnb achieves 90% of the market but is still driving down prices while maintaining standards. It is an interesting point because it seems to me that monopoly is almost built into the digital environment. It is unthinkable that Google’s market share will drop from 94% to 40%. Things do not work that way. How do we cope with it? Is regulation the answer to ensure that the monopolies are acting in the public interest?
Robert Colvile: If I was locked in a room for three months and told, “Go away and solve one public policy issue”, this is what I would devote myself to. It is utterly fascinating and vitally important. Amazon is the best possible example of this. To some extent, Amazon’s stratospheric share price is predicated on eventual monopoly. People are effectively making a bet that it will grow and grow and swallow more and more markets. In the process it will deliver enormous value to consumers, who absolutely love it. It will put the squeeze on quite a lot of producers, as Walmart and Tesco did in their day. It is almost more powerful. In the old days, if you were a retailer you could get space on the top shelf in the supermarket. Now, no one scrolls on to the second page. The result that comes up at the top of Amazon is almost certainly the thing that you buy. Likewise, the result that Alexa gives you when you say, “Alexa, order me some milk” is the thing that you buy. If Amazon has its own brand of milk, at that point that is an incredibly powerful proposition. Yet, it is delivering benefits. It is a tricky thing to untangle.
Ben Thompson writes Stratechery. It is extremely good. One of his suggestions, for example, is that no one with a dominant position on one platform should be allowed to buy another platform. Facebook should not have been allowed to buy Instagram and WhatsApp because that has entrenched its position. We need to start thinking about indirect monopoly. Even when the domination of the home market is not anti-competitive, it generates enormous profits that can be used to buy other companies, invest in other areas and expand. Google is doing self-driving cars and AI, which is all funded by the core advertising monopoly. The problem we have is not that these companies have done anything wrong. It is that they have done everything right and they are in markets which are structured to deliver outsize rewards to the people who do that.
Lord Gordon of Strathblane: Do your colleagues want to add a comment? If not, I will follow on with another question that was addressed by a witness in a previous session. Some companies are almost geared to being taken over by one of the biggies. A box is ticked—they are happy. Is there a public interest that should be unhappy? Is there a consumer interest that should be considered, as well as simply the interests of the company being taken over? Is it another three years or another three months?
Robert Colvile: No—it is about three seconds: if you founded a company you can do what the hell you want with it, unless it is illegal.
Lord Gordon of Strathblane: Sorry, you cannot. If it is in the offline world and that company becomes too big, public interest tests start to apply, particularly in your former career in the media. If Rupert Murdoch had anything like the dominance in the newspaper industry that Amazon has in the retail—
Robert Colvile: Yes. That is for government, Andrew Tyrie, the CMA and all the other people to decide. If I start a company and it gets big and Google wants to buy me, yes, the Government have the right to say, “No, we believe this will be harmful to the competition”, but you cannot say to me, “No, you should not be seeking to sell your company”. That is a very weird position to get into.
Lord Gordon of Strathblane: It may well be. You may be right. If you ran a local newspaper, for example, and wanted to be taken over by a larger local newspaper nearby, the Competition Commission would prevent it. Maybe it should not prevent it, even in the offline world, but can we live with the two worlds coexisting with different rules?
Jamie Bartlett: Mega-tech monopolies in the next 10 years will do incredible harm to democracy but not to consumer welfare. It would be brilliant for consumer welfare but not for the health of democracy. That is the crux of the problem. The monopoly law, based certainly in the US, is obviously based on harm to consumers and particularly prices. Robert is right. It is a fascinating area—working out how you redefine a monopoly if it is not about consumer welfare. While it is potentially a proportion of a share in a very small market—i.e., online advertising—you could say that Google and Facebook in online advertising is something like a duopoly. The proportion of data that you have on users is another issue that people have discussed, as well as cross-industry diversification, as you have mentioned. No one has the answer to it. Many people have tried to work this out. There is not an obvious answer.
The reason I am worried about the future is that the nature of both the internet of things and AI means that the trends that have taken us to Google or Amazon are going to be accelerated for the same reason. It is definitely one that you have to be very careful and worried about. I am not entirely sure I have the answer. If you split Google into two, it would be a less effective and less efficient company. It would not be as good. It is based on having the amount of data that it has.
The Chairman: Mr Laybourn-Langton, is that anywhere the answer?
Laurie Laybourn-Langton: No. We have to break it down by certain functions. The penetration of platforms and apps into transport in an urban environment is vastly different from search engines and other products that Google gives. There is still a response that can be done in particular activities. It would be interesting to explore areas around regulation when you enter certain markets. I believe Facebook has had banking licences in the United States of America for a number of years and is conferred with an enormous competitive advantage because of the data that it holds on many people. It may be that in entering that market officially you would have to, for example, ensure that it has open data so it fits with the open banking movement.
The third point I would make is that having a much more mixed economy of the ownership models around certain platforms could be quite interesting. When it comes to platforms entering these new technology markets, why do we not have state investment banks in the UK that would enable the public sector to invest, to direct technological development and to regain the value? We must remember that many of the technologies that underpin this were developed by state investment in the United States.
The Chairman: Has your point been answered Baroness McIntosh?
Baroness McIntosh of Hudnall: This issue about the undermining of democratic process, which is implicit in some of the answers that particularly Mr Bartlett has given, is extremely important. We have let it go past. It goes back to something that Mr Laybourn-Langton said earlier about relying on democratic process, in a way, to try to create new ways of thinking about how to regulate. If what we are trying to regulate is in the business of undermining the very processes that we are relying upon to deliver that regulation, we are in a vicious spiral. Is there some way we can think about this that does not invite us to go outside and cut our throats? This is very, very challenging. The slightly despairing tone of Mr Bartlett’s voice on a number of occasions should not be allowed to go by unnoticed, as it will not show up in the transcript.
The Chairman: Mr Bartlett, can you help us or not?
Jamie Bartlett: Could I get back to you in writing?
Baroness McIntosh of Hudnall: Yes. It would be very helpful.
The Chairman: You have been very thoughtful on this, and further thoughts in writing from any of the witnesses would be very welcome.
Q56 Lord Gordon of Strathblane: I have a relatively simple question. We are about to leave the EU, where most of the regulation of the internet takes place. How soluble is the problem that that creates?
Robert Colvile: For me, GDPR is a perfect example of why we are leaving the EU. Equally, we will still be bound by it after we leave. It goes beyond the EU. Jamie mentioned China earlier. The reason the internet has grown and succeeded and has been pretty brilliant, for all its problems—there is enormous happiness, opportunity and wealth—is that it has been a global, open thing that stretches from country to country. One of the difficulties for lawmakers in the UK is that if we go out on our own and say, “We are going to do all these different regulatory things from everyone else”, and suddenly the EU starts saying, “We are going to do something different from America”, suddenly everything starts to balkanise a bit. I believe three of the 10 largest internet firms in the world are in China. That is a very different model of the internet. It is a much more regulated and state-dominated one. The more the internet balkanises, the more there are different regulatory regimes here and there, the more you lose the connectivity between countries that gives us strength.
Jamie Bartlett: I disagree with Rob on the GDPR. It is a good piece of legislation.
Robert Colvile: You do not run a small business.
Jamie Bartlett: No, but I worked for one. Some of my despairing tone might be slightly helped by some of that. What we are doing in the UK in relation to GDRP, which is basically having it, is a reasonable approach. We will probably try to stay quite closely aligned to what the European Union does on this, and that is probably the right approach. Another area that is very important is that the UK continues to work with the EU when it comes to some of the other colloquial bodies, such as ICAN and the Internet Governance Forum, where a lot of these problems about the fragmentation of the internet are currently unfolding. It is quite important that we work together with other democratic countries on how the internet is regulated.
Laurie Laybourn-Langton: On this particular point, there are costs and benefits of leaving the EU, potentially. If we break away from the power of a regional bloc that has at least tried to resist some of the major platform giants in the world—in which China, at first view, does not necessarily agree with us on certain ideas around liberty and what these platforms should be doing—that could potentially be a very large cost. Alignment with a lot of what is being developed in the EU is probably a very important thing.
I would like to inject a note of optimism in response to Baroness McIntosh’s question. We would be particularly interested in ensuring that we slightly break the narrative. It is very important that we see these massive giants like Facebook and Google and we are fixated on them and their activities, rightly so, but we must remember that there are other markets that platforms are penetrating into, or are already in, and others that they will penetrate into. Within that context, we would advocate a much more mixed approach whereby other non-monopolists can be encouraged within those markets. That would create a much more mixed economy model across the piece, enabling us to be less focused on search and other areas which are very dangerous and well-developed into monopolisation, but would also ensure there is a slight counterbalance. One could imagine a situation where you are able to develop some municipal tools for certain regions in the UK that could support the development of democratic engagement that could work as a counterpoint to the power of some of the monopolists as well.
The Chairman: We have run out of time. I am going to take one final question. A number of the members of the Committee are trying to catch my eye. I am going to ask them to let me know their question afterward and we will forward it to you in writing.
Q57 The Lord Bishop of Chelmsford: A lot of this has been touched on in the previous answer. I want to ask about new entrants into the marketplace. How can they be enabled to compete with the established platforms, given network effects and some of the other things you have said?
The Chairman: You have touched on it, but if there is anything you would like to add on that—we do want to think about the whole economic impact. Are there any further thoughts you have on entrants into the market?
Robert Colvile: If you wanted to enter the search engine market, you would be an idiot. There are some markets where it is not going to happen. Partly, you hope that there is a chain whereby technology develops as it did previously. Every time we worried about previous dominant companies, a new paradigm came along. There is a counterintuitive issue here, which is that the platforms themselves do spur enormous innovation. AWS, Amazon Web Services, for example, or even the Amazon sales platform itself, have been an immense boon to small and large companies that can use it. You can now literally start a company processing vast amounts of data from your bedroom because AWS will rent you the software power to do it. Whatever we can do to encourage competition and new entrants, we should be doing, which is one of the reasons why I object to the GDPR. Its version of data portability in social networks, for example, is such that it effectively makes it official that no one will ever migrate their social network again. It locks people into Facebook because of the way the provisions work.
Jamie Bartlett: A very specific and simple one is the continued and maybe accelerated efforts by government to make more data that is machine-readable, especially with internet of things devices, transportation data, urban data. Government are one of the largest producers of data. The more they can do to make that available, the more other companies will be able to use that and compete. We are sitting on incredible amounts of data. The NHS is probably the most incredible source of data for medical start-up companies. Anything we can do to encourage that would be a huge positive. That is my optimistic tone.
The Chairman: That is a very interesting point.
Laurie Laybourn-Langton: I would echo that and say that it is potentially a shame that in the development of products using NHS data, we will not have control of what Alphabet DeepMind builds off the back of that. There could have been an alternative situation whereby, by centralising NHS data, the NHS could have had a leveraging position over those who create the products and owned them itself, to ensure that they were developed by other non-Alphabet developers as well. In doing so, that could be linked to the embryonic industrial strategy as well. That is all about determining outcomes in the economy. Outcomes in the economy will be increasingly determined by platforms. Why not have some kind of direction to the development of those platforms?
Baroness Quin: I am struggling somewhat with the statement by Mr Colvile that the GDPR is a perfect reason for leaving the EU, but we are going to be bound by it anyway.
Robert Colvile: I voted remain. I do not like the GDPR.
Baroness Quin: I would like to reinforce the point our Chairman made—if you could send us in writing your thoughts about the consequences of Brexit, the challenges and the possible opportunities resulting from that. Similarly, please put in writing to us your thoughts about the international dimension for the future: whether there are international bodies we ought to be trying to increase involvement with, or whether there is a gap for new international bodies—particularly given the points about China and so on—and how you see that international dimension carrying forward into the future.
The Chairman: We are going to ask you to respond in writing to that. We have had a very interesting session. You have given some perspectives on the subject that we have not had from previous witnesses. Of course, there are some concerns, but we have had quite a bit of clarity on some aspects of regulation that we have not had before and we welcome and appreciate it. The clerk will write to you and ask you to respond to Baroness Quin’s questions on Brexit and to some further questions that we would have liked to ask on personal data and how it is handled. If you would be kind enough to respond to us either specifically or by sending us other material, we would appreciate it.
May I thank our witnesses for their time today? It has been illuminating. We would have liked two hours with you, so you can maybe come back some other day. Thank you very much indeed.
