final logo red (RGB)

 

Select Committee on Communications and Digital

Corrected oral evidence: Freedom of expression online

Tuesday 9 March 2021

4 pm

 

Watch the meeting

Members present: Viscount Colville of Culross (The Chair); Baroness Bull; Baroness Buscombe; Baroness Featherstone; Lord Gilbert of Panteg; Baroness Grender; Lord Griffiths of Burry Port; Lord Lipsey; Lord McInnes of Kilwinning; Baroness Rebuck; Lord Stevenson of Balmacara; Lord Vaizey of Didcot; The Lord Bishop of Worcester.

Evidence Session No. 16              Virtual Proceeding              Questions 139 - 146

 

Witnesses

I: Jim Killock, Executive Director, Open Rights Group; Julia Reda, Researcher, Society for Civil Rights, and former MEP.

 

USE OF THE TRANSCRIPT

This is a corrected transcript of evidence taken in public and webcast on www.parliamentlive.tv.

 

 


14

 

Examination of witnesses

Jim Killock and Julia Reda.

Q139         The Chair: Good afternoon and welcome to our second session today. We are joined by Jim Killock, executive director of the Open Rights Group. He specialises in working on data protection and privacy issues. He sits on the governance board of CREATe, the UK’s research centre for copyright. Joining him is Julia Reda, a former Member of the European Parliament and a former member of the Pirate Party. Now she is a copyright reformer working for GFF, the German Society for Civil Rights, where she coordinates the control project, defending the freedom of communications. This session is being broadcast live online and a transcript will be taken.

Julia, would you like to introduce yourself and tell us a bit more about the work you are doing to reform copyright and defend freedom of speech? It would be useful for the committee if you give us a perspective on the broad themes we need to consider when looking at how copyright law and its enforcement are relevant to the freedom of expression online.

Julia Reda: I am Julia Reda, not to be confused with Julia Reid, who is another former Member of the European Parliament. I use they/them pronouns and I was a Member of the European Parliament from 2014 to 2019. I was one of the foremost critics of the EU copyright directive that was being negotiated at the time, which is now being transposed by the member states of the European Union into national laws.

The reasons why I opposed this directive are quite different from some of the freedom of expression issues that we have heard on the previous panel. The copyright issues that arise with online platforms are not about whether platforms should be allowed to give themselves rules to protect certain communities, such as rightsholders or the trans communities, and then enforce those rules. They are about a law that would require platforms to automatically remove copyright infringement in a manner that would also lead to the removal of non-infringing content in the process.

At the core of the copyright and freedom of expression problems we are seeing, I believe, is the use of automated removal of content by algorithms. That is fundamentally incompatible with the nature of copyright law, which is dependent on context. For example, the right to quotation and the right to parody are exceptions from copyright that are really fundamental to the protection of freedom of expression, or the use of content that is in the public domain. All these types of perfectly legal expression are falsely removed by automated copyright enforcement through algorithms.

The Chair: Thank you. That was very interesting. Jim, would you like to introduce yourself and give us the broad themes we need to consider in this area, please?

Jim Killock: I am the executive director of the Open Rights Group. We have worked on copyright for the whole of our 15 years. We campaigned successfully for an exception in UK law for the right to parody in copyright. I agree entirely with Julia that the enforcement issues, particularly automated takedowns, are really important. You also have to think about the notice and takedown procedures around copyright infringement, but it is worth remembering that copyright starts as a restriction on expression. We accept it as a society because it is necessary for people to receive remuneration, but in essence it is a restriction on the dissemination of material, so it is always going to have some free expression impacts. It is, therefore, very important to consider where those lines are drawn.

Down the years, we have seen impacts on non-commercial uses. We have seen large amounts of even classical music removed from YouTube, because it is misidentified as particular commercial performances when it is non-commercial voluntary performances. We have seen political material and campaigns such as Greenpeace’s Darth Vader video criticising Volkswagen taken down after claims of copyright infringement. We have even seen people cheering football goals on Twitter being removed under claims of copyright infringement.

Clearly, none of those is intended by these systems, but they happen in practice. The other thing to remember here is that the UK does not really have any legal framework for this, so at the moment we de facto lean on what the United States provides, rather than using the systems we have in law.

Q140         Baroness Rebuck: Thank you very much for those introductions. You have started to get into the heart of the matter there. I am not a lawyer, but my career for the past 40 years has been as a book publisher. Needless to say, we respected copyright and saw it as a way of safeguarding freedom of expression, to enable writers to have time to think and research, and novelists to develop stories that could go on to change people’s lives and be a source of material for the wider creative industries. Only yesterday, a Washington think tank debated the value of storytelling in strengthening democratic values.

We fought to maintain copyright, as I saw it, to keep the pipeline of ideas in the humanities and the sciences alive by fairly remunerating authors. That is the legacy, which meets the modern world of online activity. My first fairly hesitant question is whether UK copyright law adequately protects freedom of expression. Jim, let us start with you, because you alluded to that in your introduction.

Jim Killock: Julia has done really good work on this, so I will be interested to hear what she says. There are some clear problems. There is not a de minimis sampling. There is not a thing that says, “If you take a small bit of a work and use it in a musical work, that is not infringement”. In fact, in the UK, you have to pay licence fees in order to do that. That seems excessive.

In general, there is no overarching free expression defence. There is a range of activities that are completely legitimate and would not affect the commercial value of works. These could be brought in through something such as a fair use exception, which was not restricted and defined, but just said, “You can do things, so long as they do not impact the commercial value of the original rightsholders. So long as you are not treading that far and you use only what is necessary, you can do what you need to do”. Something such as that would cover many of the circumstances in which it is currently quite hard for people to ask, “Could I use a bit of this?”

The obvious example is the way people create memes. Many of these would fall into parody defences, but they would not necessarily always be regarded as a parody. Those are the kinds of examples where copyright law is not currently protecting free expression adequately.

Baroness Rebuck: Julia, would you agree that something such as the American system of fair use would level the playing field?

Julia Reda: Yes, I agree with Jim on that point. When the European copyright directive was being discussed, there was a proposal on the table to introduce user-generated content exceptions along the lines of US fair use. It was very narrowly defeated in the Committee on Legal Affairs. The main argument against it was that fair use was a common law concept that would not fit very well into the continental European legal tradition, which I do not agree with, but it is perhaps even less of an issue in the UK context. It would be a good idea to look at fair use as a possible improvement of free speech protection in the UK system.

I would like to highlight, though, the exceptions that already exist in UK copyright law. In 2014, there was the introduction of fair dealing for parody, caricature and pastiche. It is very difficult for internet users to make use of that exception for a number of reasons. On the one hand, it is because its boundaries are somewhat unclear, but it is also because there is no real recourse to say to the platforms, “Actually, this is a legal parody”.

To give you a practical example, I am sure some of you will have heard of the Marsh family, who have been doing parody videos on YouTube during the pandemic to cheer people up. They have loosely based a lot of their parody songs on popular music, but one of their videos was taken down for a period, until Sony eventually agreed to have it put back online.

This song was called “Have the New Jab”. They were using the melody of Leonard Cohen’s “Hallelujah”, but were playing their own instruments and had written their own lyrics. I believe that this could very well qualify as fair dealing, because they were clearly putting a lot of their own work into it. They were not detracting from the commercial success of the original—on the contrary, I would argue. They were doing all of this for a good cause.

The problem is that there is not really a process to tell YouTube that its copyright filter should keep this parody online, because all that the filter can really distinguish is whether the melody that is protected by copyright is present in the upload. The technology is not capable of analysing the context and finding out, for example, whether something is a parody or a quotation.

Baroness Rebuck: I take that point. Would you agree, if you look at it from the rightsholder’s point of view, that asking the platform, which is very large, very wealthy and controls the algorithm, to do some more research and finessing of the algorithm would avoid that kind of takedown, which was agreed to be unnecessary by the record company at the end of the day?

Julia Reda: In some cases, an improvement of the algorithm would help, but in other cases the problems are insurmountable. An example of where it would help is the classical music that is quite often being erroneously blocked on platforms such as Facebook or YouTube. It is simply an error of the algorithm, which does not adequately distinguish between the copyright in the classical music itself, which has expired as we are usually talking about works by composers from the 17th century, and particular sound recordings that are protected by neighbouring rights.

If a record company perfectly legitimately says, “I would like copies of this record not to be online”, and then the platform, because the algorithm does not make this distinction, takes down other recordings of the same song, the platforms would be able to remedy that. But an algorithm is never going to develop a sense of humour and be able to distinguish between a copyright infringement and a legal parody, because the algorithms look for patterns that are the same and do comparisons. They are not capable of understanding the context in which a work is being used.

Jim Killock: If algorithms develop a sense of humour, we will have much bigger problems than copyright to worry about. Algorithms are a blunt instrument. At the moment, Facebook requires the rightsholder to intervene and to say, “Yes, this identification is correct. I can say this is my music”. It is clear that the rightsholders, when they do that, are paying very little attention to what they are looking at. They are just saying, “Yes, that person playing a violin is playing something that is my record”. They are not looking; they are just ticking the box. They are basically gaming the system. They get revenue out of that. They are not paying enough attention to it, because they are making mistakes. It is human review, ultimately, that is making decisions to take down.

They get revenue out of it, so there is a perverse incentive for them to claim copyright, even when they do not have it, but also there is no penalty to it. For instance, in the UK, there is a right for notice and takedown in defamation. The user can put in a counternotice, but people start filing notices they know are entirely inappropriate. In the case of YouTube, people take down material just because they can, they are worried about it and they do not want to make a mistake that somehow impacts their revenues. It is easier to take down.

You need a penalty for people who abuse the right that they are given to remove material. When we start to look at these systems, we need to ask, “What is the right incentive for companies or others who are taking material down to make sure that they have good behaviour?” Without that, we will get bad behaviour, as we are seeing.

Baroness Rebuck: Thank you. That is a good point.

Q141         Lord Vaizey of Didcot: It is good to see you again, Jim. We came up against each other when I was a Minister, discussing issues such as piracy. It is nice to meet you, Julia. I wanted to ask about your views on the European copyright directive, which has been in place for almost two years. I would love to know how you think it is working.

Julia, without wishing to put words in your mouth, you said earlier that, if a record company quite legitimately asked a platform not to put a record online, that was legitimate. I wonder whether both of you accept the principle of what the European policymakers decided was the right position, which is that copyright holders should be able to ask for takedowns. We have previously been discussing the mechanics and whether it can work with an algorithm. I love the idea, by the way, of algorithms developing a sense of humour. I was terrified that they would kill us, but the bad jokes that would proliferate if they developed a sense of humour would be even worse. Is the principle right, Julia?

Jim, I will not dare assess what your view is on Brexit, but it has given us a pass on the copyright directive. Indeed, when the directive was passing, we had the very odd position of civil rights groups and big tech, which is now a villain, campaigning together to stop what was then Article 13. Julia, how is it working and what do you think about it? Jim, should it be part of British law? That incorporates the question of how it is working and what you think about it.

Julia Reda: I will start with the question as to whether record companies should be allowed to ask for takedown of copyright infringements. Yes, I agree they should, but it is important that those requests for takedown have to be specific to an infringement. The problem with the copyright directive is that it is a much blunter tool. It makes certain commercial platforms directly liable for copyright infringements that they did not commit.

They become liable for copyright infringements that are committed by their users. In order to escape that liability, they have to try to get licences for all types of copyrighted material, even material that they may not have any business interest in being on the platform in the first place. They also have to ensure that certain copyright-protected works cannot be uploaded at all. The problem is that, if a rightsholder simply says, “Here is a work that I hold the rights to and I do not want it to be uploaded”, that would also lead to the blocking of legitimate uses of that work.

In effect, we here have a directive that is requiring platforms, in order to be protected from liability, to use filtering mechanisms that will lead to the blocking of legal content. There, we have a real problem with freedom of expression, because requiring a platform to block something that is legal and where no court has come to the conclusion that an infringement has taken place is a form of prior restraint, according to the case law of the European Court of Human Rights. This is why also the European Court of Justice is currently examining Article 17 of the copyright directive on whether it is compatible with fundamental rights, particularly the freedom of expression.

At this point, no member state in the EU has started enforcing this law. The directive was passed two years ago, but the implementation deadline is in June 2021, so member states still have a couple of months to implement this law and it has proven extremely difficult. Some member states have resorted to simply passing a law that empowers the Government to implement this provision by decree without giving any guidance on the fundamental rights issues, which I find very concerning.

Other member states are struggling to somehow reconcile the contradictory provisions of Article 17, because not only does it say that platforms have to block illegal uses of copyright-protected works, but it rightly says that legal uses must not be blocked in the process. It is, in a way, a cop-out by the European legislator not to tell member states how they should achieve these two conflicting goals, because there simply is no technology capable of perfectly distinguishing between legal and illegal uses of copyright-protected works.

Lord Vaizey of Didcot: That is a fascinating answer. Thank you very much. It would be great to get some more information. It is something the committee should look into.

Jim Killock: Thanks for the introduction about the work we were both involved in with the Digital Economy Act 2010. There is a big parallel there. You, as a Minister, were extremely wise in managing to delay that. I see it has not been revived since, so whether we were really in conflict on that I am not so sure. As an upshot of that, you see there a law that was passed that was dead on arrival. It was unworkable and it was just a question of time until people realised that they had been sold a very poor approach by rightsholders.

The same happened in the EU copyright directive. It is a blunt instrument. It is poorly thought out. All the attempts to introduce balance were swiped away by rightsholder organisations that wanted the maximum result possible. As a result, we have something unworkable. It is worth sitting back and seeing how unworkable that is, and then thinking about the Digital Services Act that is being looked at in the EU and whether those provisions around notice, takedown and appeals systems might be a better model. That way, there is a balance between user rights, rightsholder organisations being able to get what they need done, and the platforms not being asked to do the impossible.

The current policy of waiting is a good one. More generally, this committee and others are going to have to start thinking about which bits of European and American legislation set a good model. As I say, the DSA is probably a better approach than the copyright directive.

Lord Vaizey of Didcot: Jim, that is a brilliant answer as well and you have pointed this Committee to a very interesting issue that could be quite a chunky bit of our final report. Thank you very much.

Q142         Lord Griffiths of Burry Port: I have been given a question and I want to see whether you agree with me, but you have probably already answered it. If you think you have not answered all aspects of it, you will tell me where you can supply an opinion that goes against my conclusion. Do platforms enforce copyright consistently and in a manner compatible with freedom of expression?

I note previous evidence given to this Committee by Jimmy Wales, who says, “In the Digital Millennium Copyright Act in the US is a very clear notice and takedown procedure. This happens fairly often at Wikipedia, although not that often”. That sentence reveals that there are various strands, nuances and incompatibilities, as we have heard about: the role of algorithms, which may or may not help in certain cases, whether they have a sense of humour and so on. It seems to me that we should not expect platforms to enforce copyright consistently while our thinking is inconsistent and when we are drawing on different models to provide us with a possible way forward.

Julia Reda: You are absolutely right in pointing towards certain inconsistencies. A big one is the attempt to reconcile national copyright laws with the international nature of the internet. I can only speculate on what Jimmy Wales was trying to say. It may have been that Wikipedia receives a number of takedown requests, but only a small number of them end up referring to actual copyright infringements that they then remove.

A big problem when looking at the requirements for platforms to enforce copyright law is the question: which copyright law? For example, Germany at the moment does not have a parody exception. Perhaps it is because we do not have a sense of humour. What if somebody in the UK uploads a perfectly legal parody, under UK law, to an online platform based in Ireland and then a Germany copyright holder requests to have it taken down? Copyright law does not have particularly satisfying answers to these questions. To the extent that it has answers in private international law, they tend to favour the removal of the content, which is a freedom of expression issue, because it limits the possibility for national legislators to allow certain types of uses for the purpose of free expression.

Even though the online platforms are very different in terms of size, whether or not they are commercial and what sort of business model they have, many of them implement the US Digital Millennium Copyright Act internationally. Some elements of that Act are worth looking at and possibly copying. The Digital Services Act does a relatively good job of picking up some of those positive elements, but there are also quite concerning elements.

For example, the Digital Millennium Copyright Act requires platforms to have a repeat infringer policy, which, on YouTube, for example, is implemented in the form of a three-strike rule. If somebody uses YouTube’s upload filter, Content ID, to have one of your videos removed or demonetised, and you are okay with that and do not complain about it, nothing bad is going to happen to you or your channel. If you try to assert your rights and say, “I have a licence for this use”, “This is in the public domain” or “I am using my right to quotation”, and then YouTube ends up deciding that you were wrong, you receive a copyright strike. If you receive three copyright strikes, your channel gets removed.

It is very damaging, for example, to an independent artist or a non-governmental organisation to not have a channel at all any more, which is why people are disincentivised from asserting their rights on those platforms. There are elements of the US DMCA that have direct effects in the UK or the European Union on the speech of citizens over here, which should not be copied in this form, because they disincentivise seeking recourse against wrongful removal of legal content.

Jim Killock: The whole process around notice and takedown is extremely important, as Julia has outlined. The other invidious situation with a lack of a legal structure in the UK around this is that UK platforms have no protection. If you want to run a website, a forum or a competitor to Facebook, you do not have the protections available. If you get a notice of infringement of copyright, you become instantly liable and you have to effectively remove it as the platform. If it is a UK publisher or a UK platform, the user has no means to defend themselves. That is one reason why the US platforms take the view that they are publishing in the US and they will rely on the US legal framework, because that is easier.

That has another consequence. If you want to contest an infringement, you have to accept US jurisdiction. This means you might be faced with going to the US courts to defend your parody, which may have no commercial value, fines being set against you or even losing the right to travel to the US as a result of having been found to have infringed copyright in the States, which can be a criminal matter. We really need to have a UK legal framework and we will need to think about that quite carefully. We will need to think about how the EU develops its framework around this.

Lord Griffiths of Burry Port: Platforms do not enforce copyright consistently, largely because there is so much inconsistency about. We are waiting for something clearer, in order to have a basis for formulating something that will work and perhaps even work across borders.

Jim Killock: Yes. Ultimately, it is not the platforms’ job to enforce anything. They make other sorts of content decisions, but with copyright it is a matter between the rightsholder and the potential infringer. For them to be making the judgments is somewhat inappropriate. We have to think of ways to do this that both work with quite a considerable number of complaints and quite a lot of abusive behaviour from some users, and still allow users to defend themselves when they are doing something that is correct and the accusations are incorrect.

Q143         Baroness Featherstone: Having just talked about copyright as a model, the clear notice and takedown procedure is well understood, but could that model provide a way for taking down other content? For example, if there was a complaint about content, could or should there be an external body to adjudicate between complainant and author? This is not about copyright, which is covered in law. This is more nuanced. If so, what might that look like and how would it be paid for? Who might perform that function? Would it be platforms or some kind of arbitration body? Assuming the model of copyright provides some form that could be copied, what is your view on whether it could be adjusted to other sorts of content? Jim, you go first.

Jim Killock: I would like to hear from Julia a bit about how the models are developing in the DSA around appeals mechanisms and so on. That would be quite informative.

Julia Reda: It is difficult to generalise about notice and takedown procedures from copyright law and other types of illegal content, because different types of illegal content require different approaches. For example, there are types of material that are illegal regardless of the context in which they are being used. Those materials, for example, should be blocked regardless of who sends a takedown notice. If we are talking, for example, about images of child abuse, it would be counterproductive to ask people who want to report something such as this on a platform to have to identify themselves. Anybody should be able to report it and it should be possible to take it down.

Baroness Featherstone: I was not really thinking of things that were so obviously illegal.

Julia Reda: You are thinking of defamation. There are some elements that are specific to copyright law, because it is not just about the context in which material is being used. There is also a lack of authoritative information about the rights status of works. Even compared to other types of intellectual property, such as trademarks or patents, there is no copyright register that a platform could consult in order, for example, to solve questions over who the legitimate rightsholder is.

An alternative dispute resolution mechanism could play a role, but at the end of the day it is important that they have the information necessary to solve those issues. If you want to enforce copyright on online platforms, a requirement to publicly disclose which rights you hold would be helpful as a basis. There are some quite interesting voluntary projects around transparency of takedowns in the United States. There is a database run by Harvard University called Lumen, where some platforms voluntarily disclose the takedown notices they receive.

These kinds of mandatory transparency measures can really help alternative dispute resolution bodies identify problems with such systems. It would not just take the commercial platforms out of the role of being the arbiter on questions where they are not necessarily neutral. It would also help, for example, researchers and investigative journalists to identify patterns of misuse of those systems.

For example, in the case of the Lumen database, it gave rise to a very detailed investigation by the Wall Street Journal, which found that tens of thousands of links to legitimate news articles had been removed from Google search results on the basis of false claims of copyright infringement. The investigative journalists were able to identify that because they had access to this voluntary database. A combination of mandatory transparency rules and an alternative dispute resolution procedure, which are both proposed in the Digital Services Act, would be quite helpful.

Baroness Featherstone: Jim, if it is more a battle of content moderation than something covered in the law, would that kind of model work in any way?

Jim Killock: It certainly can help. It is perfectly valid for there to be some kind of dispute resolution mechanism or appeal systems. The Facebook model is currently extremely inadequate, because a handful of items are being looked at and it cannot look at account suspensions, unless specifically requested by Facebook. It is only content questions. I believe the Facebook Oversight Board is trying to do the right thing. They want to get under the hood, as they told you at your last meeting, but they are a long way from being able to hold Facebook accountable.

If platforms are monopolistic and we are thinking about more general moderation issues, as you were discussing with the previous panel, the way to improve the experience for everybody is to think about social media diversity. Why are there only two platforms, not many? Why is it that I have to use Twitter to interact with other Twitter users? I do not have to have a Vodafone phone, a BT phone, an EE phone and an O2 phone to talk to people across those networks, but when it is social media we do. That lack of interoperability creates the lack of social media diversity, because you have to be on the platform in order to use it, and then you have to accept the experience and the content moderation rules. Everything becomes a terribly blunt instrument. The experience becomes extremely arid in terms of your ability to tweak how it works and to get the content you want.

If we had competitors that were moderating and prioritising in a way that users felt was reasonable, we would have competition on user experience and we would be able to solve these problems at least partly through the market. That is not to say that regulation is not needed, but competition policy, just as it is in traditional media, is extremely important to protect free expression. We must not accept the idea that monopolies are just going to be governed through regulation and that will solve all the free expression issues, because it will not, no more than it would if we had two newspaper proprietors.

Baroness Bull: Is anybody, secretly or not so secretly, working on the idea of interoperability between the platforms? In my fanciful world where there is a publicly owned digital space, it would be open to all, but it does not exist. I just wondered whether there is anybody who is secretly in a dark room somewhere scheming around this.

Jim Killock: Julia may be able to fill us in on what is happening, not so secretly, in the European Union context.

Julia Reda: The European Parliament did an initiative report requesting certain elements from the European Commission in the Digital Services Act. I believe it includes a call for interoperability, which unfortunately the Commission has not yet picked up. There is now room for improvement in the parliamentary process. A coalition of civil society groups in the digital rights field, led by European Digital Rights, has put forward proposals for interoperability to be included in the Digital Services Act or the Digital Markets Act.

The Chair: Thank you very much indeed. That is really helpful.

Q144         Baroness Buscombe: My question brings us on to talking, I hope, a little more about the EU Digital Services Act and how well it protects freedom of expression. I wonder if it is important that we explain for those viewing and listening to this that the Act is intended to create an EU-wide uniform framework for the handling of illegal or potentially harmful content online, the liability of online intermediaries for third-party content, which is really important, the protection of users’ fundamental rights online, and bridging the asymmetries between online intermediaries and users.

You have both said that there are some upsides to the Digital Services Act. Can we talk more about that? If you want to take this point of interoperability a little further, is this something we should be pressing for more in public policy terms?

Julia Reda: On the question of interoperability, it would also help with the problems that have arisen from the advertising-based business model of social media companies such as Facebook. From a user perspective, one of the problems with the way that Facebook is doing business is that what I see as a Facebook user is based not on my own preferences, but rather on the preferences of the advertiser.

Requiring interoperability would allow third-party companies or organisations to create alternative user experiences that I could choose from. For example, if I do not want to have a timeline based on what the algorithm thinks will keep me on the platform, but rather I say I would like to see postings in chronological order, the way it used to be, interoperability could allow that. It could help immunise us against some of the harmful effects of the advertising-based business model, in terms of where our attention is being drawn on platforms.

One positive element of the Digital Services Act is that it maintains the basic principle that online platforms are not directly liable for the actions of their users, unlike the EU copyright directive. This is very important for freedom of expression, because at the end of the day making a platform directly liable for actions of a third party incentivises overcompliance. It incentivises the platform to remove everything that looks like it might be illegal, based on very broad automated decisions.

The sheer number of postings made on social media would force platforms to make automated decisions about what is allowed and what is not. That will always lead to overremoval of content and will especially affect certain marginalised groups. For example, to the extent that platforms already voluntarily do automated content moderation, they tend to overremove accounts that write in the Cyrillic alphabet, because they might be misidentified as Russian bots.

The Digital Services Act also includes a relatively modest application of fundamental rights to the platforms. Generally speaking, fundamental rights bind only the state and protect citizens against actions of the state. But the Digital Services Act says that, while platforms are allowed to set their own rules, which is a good idea, as I do not think every platform should be required to allow everything that is legal, they have to take fundamental rights into account when applying their private rules.

This means you can make a platform that is deliberately aimed at children and has different moderation rules from a platform aimed at political discourse by the general public. You can have a platform that allows only cooking recipes. These are commercial decisions or perhaps decisions to foster a particular form of discourse, which is perfectly legitimate and speaks to Jim’s point about diversity of platforms, but the rules that are put in place should be transparent and enforced in a non-discriminatory manner, and there should be a recourse against them. For example, if a posting gets deleted, you should have a right to know why this happened and you should be able to appeal against it.

Baroness Buscombe: That is really helpful. Thank you very much.

Jim Killock: On interoperability, the Competition and Markets Authority is starting to look at that question in the UK. I hope that it is not shouting into a void. It needs some help to move along with that, because it has had pushback and has not prioritised it quite as much as people were hoping, but it is a live question.

Julia was talking about the advertising market. Another way of characterising the social media market is as an attention market. Basically, they are after our eyeballs. They want us looking at the platforms all the time. That prioritises a certain sort of experience. For want of a better word, it is like living in a tabloid universe, except it is our social universe. The same sort of content rises to the top in a social media world as in a tabloid world, and it is of similar value sometimes. It is cat videos and horrific arguments that get prioritised. That is not necessarily a great thing and competition would help us start to unbundle that.

Julia has answered the question about where the Digital Services Act tries to protect free expression. The key thing for the Committee is that it is trying to do that from the get-go, whereas the online safety Bill is trying to build that in after deciding the particular approach around online harms. While it is welcome that the Government have started to shift on that, it might have been better to start from the point of view that the whole area needs to be looked at, both as a set of content problems and as problems with free expression and user rights. I hope the Government will get there.

We have to think about whether these rules, if they get imposed in Europe, are going to be de facto what happens in the UK. You want consistency, so there is quite a tricky political question there about which bits get adopted and where we have appetite to do that.

Baroness Buscombe: Those questions are going to become quite pressing, because we are where we are now. Are we in a void or a hybrid situation, if we are not careful? Thank you both very much. That is very helpful.

Q145         Lord Stevenson of Balmacara: My question has probably already been asked. On the basis of your suggestions today, which I have found very helpful, I wanted to ask for examples of laws elsewhere that we might want to put into the exciting new law we are going to publish. I am going to ask a which I have never dared ask before, and I want Gail Rebuck to close her ears. Why is copyright so long and is a lot of the problem we have to do with that? If I had time, I would ask this supplementary, but you do not need to answer it. Is patent law, which is effectively 20 years, within which forced commercial terms are available towards the end, not a better model?

Jim Killock: I am going to speak as a veteran of the debates about extending copyright on music. You are right to say it is horrifically long. There is no good commercial reason for it. It deprives a lot of people of the use of culturally valid works, as well as creating international inconsistency. It was this year that George Orwell’s works fell out of copyright, yet Nineteen Eighty-Four is the property of us all as a phrase and a cultural reference. It seems a terrible shame that we have not been able to use that cultural reference much more widely and without the fear of copyright interfering over the last few years. It is still commercially impossible to do very much with Orwell’s works, because they are in copyright in the States and that is going to be the case for a while.

It is inappropriately long. If you want an answer as to why, it is successful lobbying and it suits the publishers. That is it. It is usually for a handful of works, because the vast majority of works are just not commercially important and could well fall out of copyright. There are ways to deal with that. Registration of works for copyright term renewal would limit what stays in copyright for a long time. That is a solution we could look at.

Julia Reda: I completely agree with Jim. The EU has tried to introduce additional regimes to deal with the problems of the overlong copyright terms. It may be useful to look at the provisions of the EU copyright directive that deal with out-of-commerce works, because, unfortunately, it is very difficult to shorten copyright terms after the fact. One strategy is to require registration of works after a certain period has elapsed, and you will find that a lot of copyright holders will not be interested in renewing their copyright. Another would be to look at provisions for outofcommerce works, so that, if a work can no longer be purchased on the market, there should be a possibility to use it legally.

Q146         Baroness Bull: I really want to debate the question that Lord Stevenson threw us, but that will be for another time. I want to draw on your knowledge of regulation around the world and ask whether there are possibilities for further cooperation on digital regulation. We heard from one witness, Dr Edina Harbinja, that she was not optimistic about this because of the difference in cultures, social mores and moral parameters, and that it would make collaborative regulation very difficult. I am keen to hear from both of you on that.

Julia Reda: I am a bit more optimistic, but these issues need a little more time to be sorted out. When we think about it, 20 years ago, Facebook and Wikipedia were only just starting out. In the grand scheme of things, we are not very far into the history of online platform regulation. I am optimistic that we will be able to come up with some sensible rules that can be applied across cultural contexts. A lot of good ideas in that regard are in the Digital Services Act proposal. It is worth, for copyright in particular, looking at the innovative solutions that have been found, for example, in Canada with the notice and notice system and in the United States with fair use.

Jim Killock: All of those are good examples. It is also in a negative sense worth looking at the imbalance in the Digital Millennium Copyright Act, as we were mentioning before. The power balance is not quite right. Domestically, there are the problems that arose with the notice and takedown system for defamation, in terms of reputation managers filing too many complaints on spec. One can learn from what has already been done, even if it is not entirely in a positive way.

The Chair: Thank you both very much indeed for some extremely interesting and very thought-provoking evidence, which we will bear in mind and I am sure will have a big place in the report. Thank you, Jim Killock and Julia Reda, for the insights and helpful thoughts about these issues.