Communications and Digital Committee

Uncorrected oral evidence: The Online Safety Act: implementation and impact

Tuesday 3 November 2026

2 pm

 

Watch the meeting

Members present: Baroness Keeley (The Chair); Baroness Caine of Kentish Town; Viscount Colville of Culross; Baroness Elliott of Whitburn Bay; Baroness Fleet; Lord Holmes of Richmond; Lord Kirkhope of Harrogate; Lord Knight of Weymouth; Lord McNally; Baroness Owen of Alderley Edge; Lord Tarassenko; The Lord Bishop of Winchester.

Evidence Session No. 5 Heard in Public Questions 93 - 110

 

Witnesses

Damian Collins OBE, former MP and Chair, Joint Committee on the Draft Online Safety Bill; Baroness Kidron OBE, Founder, 5Rights Foundation; Professor Lorna Woods OBE, Emeritus Professor of Internet Law, University of Essex; Rt Hon Sir Jeremy Wright KC MP, Member for Kenilworth and Southam and former Secretary of State for Digital, Culture, Media and Sport.

 

 USE OF THE TRANSCRIPT

  1. This is an uncorrected transcript of evidence taken in public and webcast on www.parliamentlive.tv.
  2. Any public use of, or reference to, the contents should make clear that neither Members nor witnesses have had the opportunity to correct the record. If in doubt as to the propriety of using the transcript, please contact the Clerk of the Committee.
  3. Members and witnesses are asked to send corrections to the Clerk of the Committee within 14 days of receipt.

Examination of witnesses

Damian Collins, Baroness Kidron, Professor Lorna Woods and Sir Jeremy Wright.

 

  1.             The Chair: Good afternoon and welcome to this meeting of the Communications and Digital Select Committee. Today we will continue our inquiry into the Online Safety Act and examine how far the implementation of the Act to date matches up with the intentions of those involved in developing and scrutinising the legislation. All of our witnesses today played an important role in that process, and I thank you for joining us today. Before we move on to questions, I shall ask you each to introduce yourselves. You do not need introduction to many of us, but please do so for the broadcast.

Sir Jeremy Wright: I was the Secretary of State at the Department for Digital, Culture, Media and Sport between 2018 and 2019, which is when we produced the online harms White Paper. Then, having left government, I chaired the All-Party Parliamentary Group on Digital Regulation and Responsibility, which acted, I hope, as a forum for several in both Houses, some of whom are around this table, to discuss the progress of the Bill and amendments to it.

Professor Lorna Woods: I am professor emerita at the University of Essex. Until recently I was professor of internet law there and a member of the Human Rights Centre, and I am adviser to the Online Safety Network. Thank you.

Damian Collins: When I was in the House of Commons, I chaired the Lords and Commons Joint Committee on the Draft Online Safety Bill, which was the pre-legislative scrutiny committee. Before that, I chaired the DCMS Select Committee and was briefly Minister for the Bill at the Department for Digital, Culture, Media and Sport as well.

Baroness Kidron: I am a Cross-Bench Peer in the House of Lords. I was also on the pre-leg committee under the chairmanship of Damian, and I played quite a big part in the passage of the Bill.

  1.             The Chair: I am sure you have all seen that Ofcom has recently published a report showing that there has been no significant change in children’s exposure to harmful content, nine months after the children’s safety duties under the Online Safety Act came into force. Given that, in your view, does the Act still have the potential to achieve its original aims, or does the framework that it established now need fundamental reform?

Sir Jeremy Wright: The answer to that is yes, it still has that potential. I suppose I should declare, as the committee knows, a sense of parental responsibility for this Act, so I am perhaps not its most impartial observer, but I think the structure of the Act will survive the test of time. We had a choice back in the day when we were trying to decide the way in which we should approach this challenge: either we could have a rules-based regulation approach, where essentially we had a list of dos and don’ts for online platforms, or we could take a different, much more principles-based approach to regulation based on a duty-of-care model. Lorna and Will Perrin, who is probably the other intellectual godfather of this particular approach, persuaded me at the time that the duty-of-care approach was right, and I am still persuaded of that. If you pursue a rules-based approach, with a list of dos and don’ts, then what you find in this space above all others is that your list becomes out of date almost as soon as you have finished writing it. So I think the approach broadly set out in the Bill is the right one.

What we have to recognise, and this is probably reflected in the Ofcom report, is that the harms multiply all the time. There are many more and many different kinds of online harms that particularly affect children, so the fact that we appear to be standing still does not necessarily mean we are failing; it may mean that the harms are multiplying and we are trying to keep pace with them. However—and you may find that this is a theme of the evidence we all give to you today—it is important that the Act is properly implemented. I think we are now in a place where the Act allows for more action than the regulator is prepared to take. A large part of the problem is that we have handed over a piece of legislation, imperfect though I accept that it is, that could be used more effectively by the regulator than it is currently.

Professor Lorna Woods: I agree with what Jeremy has said. Maybe there is a certain amount of bias, given my work on why the duty of care would be a good thing. Focusing on the choice architecture of the services reflects the fact that they are not neutral as to content and behaviours. It is also something that the companies control, so to me it seems legitimate and appropriate to link that responsibility with some responsibility for doing something about it.

I had assumptions, which perhaps proved unfounded, as to how far the regulator was prepared to go without an explicit steer in the legislation. You can see in the Ofcom codes that there is a focus on aspects that were specifically mentioned in the legislation. For example, there is a lot in the codes on the requirement to have an effective takedown system. Illegal content codes for user-to-user services and highly effective age verifications are two specific things that the Act required, and it is noticeable that they are the two specific things that Ofcom has focused on, so maybe the regulator was at sea.

To pick up on a couple of points that Jeremy made, the world has changed. Having a principles-based approach and an approach that focuses on requiring services to look at what they are developing helps with future-proofing. For that reason, I do not think that moving away from the approach we have at the moment would be right. The Online Safety Act could never have been a silver bullet; it looks at a particular segment of the market and its particular role in an environment. You are going to have to look at people posting content and the rules relating to that as well, and how those fit in.

To be the voice of pragmatism: if you try to bring in something entirely new, that disincentivises services from complying with what we have now, and then there is the time lag. It is difficult to assess how well the regime is functioning at the moment, partly because there was the time lag in the codes coming in. Then we need to wait for enforcement, and we are only really at the beginning of seeing that enforcement.

Damian Collins: Chairman, the premise of your question is that the Act has not delivered the results that were hoped for and therefore the question has to be why that is the case. I do not disagree with what Lorna or Jeremy have said, and I am sure during the course of the afternoon we will not disagree that much, but I will say that at the heart of this debate is a problem in the way that Ofcom is interpreting the powers that it has been given. Clause 10 of the Act deals with the safety duties relating to illegal harms, and you have to look at Ofcom’s written evidence to this committee and use your judgment as to whether it has interpreted the Act in the way that it is written and the way that Parliament intended at the time, specifically regarding the extent to which platforms have proactive obligations to scan for and mitigate the existence of illegal content. In the written evidence to the committee, Ofcom says that those do not exist, but the Act clearly says that there is a duty to prevent access.

Canada has recently introduced its own version of the Online Safety Act, the Safe Social Media Act, which expressly says there is no obligation to proactively scan; it is written in the Bill. It was always intended that the illegal safety duties in the Online Safety Act would give guidance to the companies. The duty of care rested on a series of priority legal offences. We expected the companies to have systems in place to address those needs. It did not require perfection, but in our view it required the companies to be aware of those obligations, and they would proactively seek to identify illegal content, act against it in order to make it hard for users to find and remove it, rather than waiting for someone to complain and then retrospectively deal with it.

In fact, during the passage of the legislation, one of the lobbies from the tech companies said that, as we were no longer in the European Union, the e-commerce directive no longer applied. In that commerce directive, the companies are protected from an obligation to monitor generally what goes on on these platforms. We did not want to introduce that in the Online Safety Act because we knew it would be used as a get-out-of-jail-free card by the companies; every time they were found to have failed in their duties, they could say, “Ah yes, but there’s no obligation for us to monitor everything”.

The way that the provisions in Clause 10 have been interpreted by the regulator is wrong. It is certainly not what we intended, and I think it is at the heart of why enforcement has been weak.

Baroness Kidron: I broadly agree with what everyone has said. The report from Ofcom that you mentioned at the outset, Chair, is a travesty. It took 10 years to get here but then, nine months after implementation, nothing has changed. I just want to put into this room the voices of desperate parents and children at the other end of that “nothing has changed” so that we do not forget about their experience as we talk about the technical aspects of the Bill.

I largely agree with my colleagues: yes, it does still work, but there are four things that have to be thought about. The first is whether the powers that have been given to Ofcom have been adequately interpreted. As you have heard, the general view on this side of the table is that they have not.

Secondly, what has changed, what has moved, what could we not have anticipated at the time and what did the Government of the day not allow us to put in? Where are the gaps, and by what mechanism will they be filled?

The third is a culture question about Ofcom: is it too cumbersome? It has a big budget and hundreds of people, but it is very hard to get a sense of urgency. Even in the recommendations there are a number of things that it could decide to do right now, but it is waiting for future action from the Government. This “slightly more in sorrow than in anger” sort of communication from Ofcom is something that the committee might look at and say, “Are they supervisory or are they enforcers?” If they are enforcers then it is on them that it is not working, or their recommendations would have been, “We need this or that power right now in order to do our job”.

The fourth issue is enforcement, where I have some sympathy with Ofcom. I do not think the enforcement powers are all there. Unless there is proper enforcement capability, the Act will not work.

One of the critical mistakes by the Government, the regulator and, to a degree, even Parliament itself was that we were all used to, and had in our imagination, regulation of sectors that are used to being regulated and willing to be so, but I cannot emphasise enough how, both ideologically and practically, the tech sector does not see itself as being willing to be regulated. When you come up with your final recommendations, it is important to understand that and see how enforcement needs to change, which we may get to later. That is critical to the way forward, and it is something that sits with all of us, not just the regulator.

The Chair: You have been very comprehensive. I am going to try and stick to the time that we have available, which means that you do not all need to speak on every question.

  1.             The Lord Bishop of Winchester: In a sense I want to pick up on your last point about a tech sector that does not want to be regulated, but I shall specifically talk about the Act’s safe harbour provision. There is some feeling that this seems to have created a ceiling rather than a floor for platforms to mitigate risk. Do you think that is right? Should the provision be revised and what, if anything, should replace it?

Baroness Kidron: I absolutely think it should be revised. There is one view of safe harbour that could mean that the regulator could see that a tech company had done all that it could have but still some harm happened; however, because it did all that it could, it was offered safe harbour. How it has been rather bureaucratically interpreted—and I would go to Lorna on whether it needed to be that way—is that they have said, “Okay, if you do everything in our codes then you are fine”, even though we have found many risks outside of the codes, and that creates an automatic gap from day one.

Sir Jeremy Wright: Can I slightly disagree with that? I think there is merit in safe harbour provisions. Because we are in a very flexible and fast-moving landscape, it is important that we are fair to those that we are regulating. If we are simply going to say to them, “We want you to keep your users safe”, but not explain precisely or at least roughly how that might be done, then that would be unfair to them. However, the safe harbour provisions need to work on the basis that Ofcom, as the regulator, has higher, not lower, expectations. I fear that what has gone wrong here is that Ofcom’s expectations are too low.

What is in the codes of practice—offering safe harbour if you do what they say—is, frankly, not ambitious enough. That is strange to me, because these are some of the most inventive companies on the planet. We can expect them to innovate all over the place in order to advance their own cause or offer us a product we are more willing to buy, so I do not see why we should not expect them to be equally imaginative and innovative when they are presenting safer options or worrying about the safety of their users. I want to see the interpretation of safe harbour being, “We offer you safe harbour if you’re innovative and ambitious and if you do as much as you proportionately can”. We might come back to the definition of proportionate because that is important, but it is important in this context to expect an ambitious response, and that is what is missing.

Damian Collins: Jeremy makes an important point: there is nothing in the Online Safety Act requiring companies to do something that they do not already have the technology to do. They do it to target advertising content at users but they do not use it to keep people safe.

One of the big structural differences between the Online Safety Act and the DSA in the EU is the fact that the safety duties concern defined harms based on laws that Parliament has already passed rather than fundamental principles of human rights. The reason for that was to help with enforceability. It was not to create a safe harbour where a minimum safety standard could be achieved and nothing more beyond it would be expected; the expectation was that it would be a floor, not a ceiling. Ofcom is already facing judicial reviews from the tech companies but we thought that, down the line, the less specific and clear the duties and what they were based on, the more open that would be to legal challenge later on. So the intention was to create certainty for the platform, in terms of both what it needed to do and what enforcement would come if it failed.

Professor Lorna Woods: I agree with what has been said. The real problem is the subject matter in the codes and the perception, seemingly by Ofcom, that it has to give specific steps rather than programmatic obligations. That is the heart of the problem. I think the regime was that companies should try their best to minimise, not that they should come up with perfection. But what they have is a rather low-level tick-box list, which has been constrained by availability of evidence and Ofcom’s concern that everything should be very precisely measurable—seemingly.

One other point on the DSA: it also envisages codes but does not have a safe harbour provision. As they are not binding, they are of a different nature to the codes in the Online Safety Act, but the DSA sees complying with codes as evidence of attempting to satisfy duties rather than an absolute guarantee.

  1.             Lord Holmes of Richmond: What implications would reform of safe harbour have for legal certainty and Ofcom’s ability to enforce the Act? How could any difficulties be addressed? Can Beeban answer first, please.

Baroness Kidron: Sorry, I did not quite understand, Chris. Were you asking about removing safe harbour?

Lord Holmes of Richmond: If you were looking to change safe harbour, what implications would that have on legal certainty and Ofcom’s abilities?

Baroness Kidron: I have always been interested in how the risk register, where Ofcom sets out the totality of risk, is much richer, much better and more straightforward than the codes themselves. It is interesting what colleagues have said, because there seems to me to be little difference between having a safe harbour that guarantees you if you have tried everything and have been ambitious versus having no safe harbour, as in not having the low bar of a code.

In spring 2025, when Ofcom was doing these codes, somebody came to me from within the companies and said, “This is ridiculous. We are going to fix feasibility at what we say we can do in 2025, and it will be there for all time”. It is a floor and it is pushing us downwards. The safe harbour is part of what makes it unambitious.

I agree with others that there must be some flexibility of interpretation like there is in the DSA. But when I went to Ofcom and said, “Could your codes be guidance and the risk register be what you are going for?”, it said, “But how do we measure that?” I said, “I do not understand how you measure against the code”. At some point, you always have to make a judgment about whether you have reached the bar. I think that the bar should be the overarching risk register. That is better.

Damian Collins: One of the benefits of the structure that we have created—and I have heard tech companies talk about why they much prefer working within the OSA than the DSA—is that you can go to a regulator and say, “If we address this problem in this way, is that going to be good enough?” That is fair enough if that leads to better regulation. What was never intended was that, if the system put in place clearly had failed and is not delivering the results expected, it can carry on failing without any intervention, when the regulator, because it consented to that approach being taken, is powerless to do anything about it. It was clearly meant as a system where risks were assessed and companies put policies in place to address those risks, but clearly that is not enough. The responsibility on the company to meet that risk should still exist. Again, the codes of practice have been interpreted in such a way that the obligations of the companies do not go beyond what is expressly called for in the code, regardless of the risks or the safety needs of the users. That is wrong; it is certainly not what we intended when the legislation was passed.

Professor Lorna Woods: I sometimes wonder why the concern about uncertainty arises here when it does not arise in relation to product safety or in relation to the Act about building standards brought in following the Grenfell Tower tragedy—I cannot remember its name. It is a similar sort of theory of putting the responsibility on the person responsible for creating or constructing, but here and here alone we suddenly see concerns about uncertainty. I do not know why this sector claims to be so different.

Obviously, when you have a new regime, there will always be questions and some bedding in, so you end up with a better understanding of rules of thumb once you have some process and experience to guide you. However, I am not sure why there should be a problem here, comparatively, which we do not see elsewhere.

Sir Jeremy Wright: I agree with all of that. When Lorna and I were discussing the principles of this way back in the mists of time, we envisaged an overarching duty of care that would be the keystone to this entire legislative edifice. That has gone, and the duty of care has been broken down into its component parts. Had we retained it, there would have been the opportunity to say, “Whatever else we say to you about detail, you’ve got this residual overarching duty of care to keep your users as safe as you reasonably, proportionately, can”.

The other point is that, as I said earlier, the regulated are entitled to an expectation—to have some idea of what this duty of care means in practical terms. If you are not expecting a good deal of people, they will default to the minimum you are asking of them. One of the problems I have with the way in which Ofcom has approached the way in which it decides what to put into the codes of practice is that it has decided that it can only recommend in a code of practice something that would be proportionate to expect a platform to do. So far, so reasonable but, when you define “proportionate” as meaning “We know it works and have evidence to show that it works”, you are actually fixing the benchmark at current good practice, because you have evidence that it works only if people have been doing it and probably doing it for long enough to create an evidence base.

My worry comes back to this point about lack of ambition. If Ofcom is saying that it will only recommend something in a code of practice, something which then gives protection from regulatory intervention—because that is what safe harbour means—and if it sets the benchmark at things it can prove work because people are already doing them, this is problematic. You can fix things at the status quo but, in a fast-moving landscape, the whole point is that we want these platforms to scan the horizon and be responsive to problems that they see emerging.

Worse than that, an innovator—a platform out there trying to do its best and move the needle on safety such that it is doing something exciting and newly effective in keeping its users safe—is then being told that it is better off retrenching back to the standard current practice, because that gives it regulatory protection. That is the exact opposite of what we are trying to achieve, which is to encourage platforms to be ambitious and to be scanning the horizon themselves, and have the weight of obligation on them through the duty of care model, so that regulators and legislators, like all of us, do not have to run to catch up all the time.

  1.             Lord Holmes of Richmond: I would like to explore, particularly with Lorna and Jeremy, your points around an overarching duty of care. There was an interesting exchange in the Joint Committee sessions in 2021 between a certain Lord Knight and the then Minister Dorries, when this point about an overarching duty of care came up. Her comments at that time were along the lines that it could not get past parliamentary counsel. She discussed with you, Jeremy, that it was not a runner. Is that your recollection? Do you believe that it would not be possible to have such a model of an overarching duty of care?

Sir Jeremy Wright: It is rare for me to disagree with Nadine, but on this occasion I do. I do think it was viable. Again, if you go back—Lorna will correct me if I have misremembered this—the original thought was that we can do something in the online space similar to the operation of the Health and Safety at Work etc. Act 1974. The Act does not require you to do something specific with the buckets on the floor or the shelving, specifically. What it says is that you have to create a safe environment. The objective here was to do the same online, and I am not persuaded that that would not have worked. As I say, it would have addressed some of the problems which we have identified already in this session. It is not a panacea. I accept that, and it is not easy to do, but it is feasible. What it would have done is set the tone for the obligation being on the platforms to do what they sensibly can to keep their users safe. I am perfectly prepared to accept, and always was, that if platforms do all that they reasonably can to keep their users safe and something bad happens to a vulnerable user none the less, I do not expect them to be in the stocks for that, but we need to set our expectations high because, again, these are companies capable of meeting high expectations. We are not asking companies to do something that they cannot do, and we should not ask them to do something they cannot do. This is asking them only to do what they can do. That should be our expectation.

  1.             Lord Holmes of Richmond: So it is not your understanding that an overarching duty of care will be problematic from the parliamentary counsel perspective?

Sir Jeremy Wright: I am going to defer to a better lawyer than I am. Lorna will tell you what she thinks but my view is that we should at least have tried. That is the kind of Bill that I would have wanted to introduce, because it is an important element that ended up being missing.

Lord Holmes of Richmond: I appreciate that.

Professor Lorna Woods: Unless someone comes up with a specific problem, I am struggling to think what the problem is because we have other sectors where a similarly broad approach has been taken. There may be problems but they are just not leaping to my mind at the moment.

Damian Collins: If you go back to the moment of that exchange, the Bill was framed and structured differently. The idea of having priority legal offences in a schedule of existing Acts of Parliament written on the face of the Bill that the safety duties will be built on was not in the Bill at the time. Actually, one of the challenges at the time was: what does the duty of care rest on? Does it rest on the Human Rights Act? Does it rest on, as it does in the EU, the Charter of Fundamental Rights of the European Union? What does it rest on? There was some concern at the time about whether we were then delegating to a regulator lots of interpretations of rules around speech, in particular, which would be where the regulator would be giving rulings on harm based on behaviours that were not necessarily unlawful.

The question at the time was very much: what does that duty of care rest on? We felt the duty of care rested on the priority legal offences, and the proper enforcement of the platform’s own terms of service. It is interesting that Ofcom is saying in its evidence, certainly with regards to younger children accessing their services, it is not entirely certain what powers it has. Again, it is another example that the way in which the powers that were given to the regulator have been used or interpreted is weaker than we thought it would be at the time.

Baroness Kidron: I am in Lorna’s school. I heard all those debates and no one gave a good reason. Maybe the useful thing to say about parliamentary counsel at the time is this: we also had an excruciating exchange about whether there was anything else online apart from content—as in parliamentary counsel could not understand that the online world was made up of data, behaviours and activities way beyond the content itself. I am sure we are going to come to it in a minute regarding some of the other things that the Act therefore failed to get at. The compromise, if you look at the Bill, was that it says that content means content and activity but it has not been interpreted as such. It has only been interpreted as content.

My reason for saying that is that it is worth the committee understanding two things. One, which is fair enough, is that it felt like novel legislation. People were anxious. You live and you learn. But something much more important is that there was an absolute failure of understanding within the department about what the technology could and could not do, and where the dangers lay. I am afraid that the voices in the horseshoe were not taken quite seriously enough at that time.

  1.             Lord Tarassenko: I am interested, possibly because of my engineering background, in design and design choices. In fact, there is an interesting blog by the president of the Royal Academy of Engineering about making the right designs for chatbots and so on, which you may have read. I think everybody in the horseshoe and in this room will probably agree that the Act has not succeeded in making online services safe by design. I have read from cover to cover the document produced by the coalition involving the 5Rights Foundation, the Molly Rose Foundation, the Online Safety Network, the NSPCC and yourself, Lorna. It has some interesting ideas but I wanted to explore a little as to whether that is one way to go. One of the things that you advocate together with that document is that the Government should introduce a statutory definition of safety by design and then require Ofcom to develop a dedicated code in the same way in which you have put a template together.

I have one question, but it is slightly different for the members of the panel today. For Jeremy and Damian, it is a general question. What would the type of proposals in this model of a safety by design code of practice require platforms to do differently from what they are doing now? Does Ofcom have the powers to introduce these different measures? For Lorna, which parts of the code could Ofcom adopt tomorrow under Section 41 and which would need legislation? Are the technical amendments that you mentioned in your document amendments to the Act? Finally, Beeban, I will give you the choice to take the general question or the specific question.

Sir Jeremy Wright: Safety by design was always the intention. That was always the guiding principle for those of us who designed this Bill from the outset. We have been disappointed to see that the focus has moved away from that towards an unnecessary focus on content itself—specific pieces of content. Of course, those are important but the design features of the websites and platforms we are talking about are a fundamental aspect of the harm that we could all see being caused and that we wish to address. It picks up exactly on the point Beeban was just making about the way in which things happen here. It is not just about whether there is bad stuff out there online. It is the fact that the bad stuff is being served up to you, particularly if you are a child, in a way that drags you towards it. It is not you going looking for it specifically. It is the algorithm and the recommender system being designed in such a way as to promote that material to you. That is a design feature. It has to be addressed through the legislation and the work of the regulator, building on that legislation. Now, I take the view—this is a constant theme, I am afraid—that there is plenty in the Act already that allows Ofcom to do that if it wishes.

Many of us worked hard to get words like “functionality” into the Act more often than at the start. That is because “functionality” covers exactly what I am describing. If you look at the way in which this Act is designed, it starts with asking the platforms to do a risk assessment, do it properly and, having identified risks, go on to address and mitigate those risks. Setting out how you go about risk assessment and what you need to look for is exactly what this Act does. Included in the list of things you have to do if you are risk assessing, particularly for material that might be harmful to children, is a list of things, including the functionality of the platform and, for example, anything about the design of the platform that keeps children on it for longer. That is rather forgotten about but it is an important part of the harm that is being done. So quite a lot of this is in the Act already.

I return to the same point. We have to be enforcing the legislation we already have. I am not against further legislation. I never thought that the Online Safety Act would be the last word on the subject. We must do the best we can with what we have first before we can make a credible argument for more legislation.

Damian Collins: There has always been a tension in the debate here. The Online Safety Act does two novel things: it establishes broad areas of intermediary liability for platforms that do not exist in America, which is why the companies have always struggled with it, and it regulates feeds.

This debate about safety by design is not an unhelpful debate to have for people who are thinking about how you regulate AI. If you are saying that you have designed a system to work in a particular way, draw data in a particular way and produce certain kinds of outcomes, is that safe or not? Content feeds, for many users, work in a safe and benign way: they could direct people towards music or films that they might like. For some people, though, there are devastating, catastrophic failures.

I think the Act should be clear that there is not a safe way to recommend self-harm and suicide content to a teenage girl. The design of the platform should prevent that happening; otherwise, it has clearly failed in its safety duties. That is how I see it. What the Act does not really address are things like excessive use—the extent to which addiction is harmful. It relies on categories of harm to help the regulator to make that decision. 

In some ways, I go back to what I said at the beginning. The regulator’s interpretation of Section 10 of the Act is partly responsible here. That section says that the safety duty is to prevent people from encountering illegal content, which is about the design of the system or platform, not individual pieces of content. It is the ability to use categories of content as an exemplar to proactively scan and identify similar content, to remove that content or mitigate it in some other way so that the user does not experience it, and to train those systems so that they do not recommend content that is illegal.

We know from other areas of platform moderation that they experiment all the time with the way the feeds work. They can upgrade and downgrade certain categories of content and where it is recommended from on the platform to change the outcome of the user experience. They do this all the time; it is well documented. I think the Act should require companies to manage the design of their platforms in such a way as to meet their safety duties.

Professor Lorna Woods: I want to say first that I agree with Jeremy. When I was thinking about this, I assumed that a duty of care would be aimed at ensuring that services were safer and that the key thing would be the way that they are designed, developed and managed. That seems to have got lost somewhere, maybe during the long legislative process, so it is important to re-establish that.

The House of Lords did manage to get in the reference to “safe by design” in Section 1, but it would be helpful to elaborate on that, not least because different sectors perhaps look at the topic slightly differently. You have “by design” discussions in data protection and cyber security, then in engineering as well. So part of the reason that we suggested that there should be a definition was to try to reconcile the different viewpoints.

When I was reading the literature, I tried to synthesise that into three. We need to think about the life cycle, as this is not just about product development; it is about how you operate and also retire products. How on earth do we retire some of this stuff safely? There is an issue of a hierarchy of control, which is often seen in engineering and in building: it is the idea that you should try to reduce the hazards before you seek to control them. For the entire product, there is a great emphasis in Ofcom’s codes—and there was that emphasis in debate when the Bill went through Parliament—on take-down, which is the end of the process. But what about account creation, recommender tools or user empowerment tools? All these things could be taken into account when trying to design safely. Coming to the UN Ruggie principles, the idea that harm should be dealt with— 

The Chair: We are getting into some very long answers, and we are three-quarters of the way through the time that we have allocated to ourselves, so could you be brief?

Professor Lorna Woods: Yes. It was just to say that the harm should be dealt with as close to source as possible. Those are principles rather than a definition. The code was there, in a way, to bolster what we have already. You could possibly incorporate principles into existing codes, but it was just trying to work with the existing structure.

Baroness Kidron: I am going to meet an engineer’s comments with an engineer’s answer. The IEEE did a technical standard for age-appropriate design; 2089 is the name of the standard. That really speaks to the work that we did in Parliament to put safety by design in Section 1. That was, in fact, the victory of the Opposition Front Bench at the time—Lord Stevenson—and we celebrated that.

So I can tell the Committee, absolutely 100%, that safety by design is the lens through which you should look at how to fix this and that there is a technical standard that would give some guidance on that.

  1.        Viscount Colville of Culross: I want to talk about a specific area of safety design, which is addiction. Jeremy, you have just talked about how the risk assessments cover a list of likely harms. That includes the question of whether this would keep children online for longer than they should be. We have had evidence from Ofcom, which said that addiction is not in the Online Safety Act; it is not a harm that is defined in that Act. Does that mean that, to get Ofcom to start chasing the problem of addiction, we need to amend the OSA?

Sir Jeremy Wright: You will not be surprised to learn that I think that it is already in the Act and, once again, that I do not agree with Ofcom about its interpretation of the Act. If you will forgive me, Madam Chair, I will refer to the Act; being a good lawyer, I carry a copy everywhere with me.

As I mentioned earlier, the Act sets out a list of things that you should do when assessing for the risk of material that could be harmful to children. There are a whole series of things in Section 11(6)(a) to (h). Some relate to content, but I want to read you (f). Again, this is one of the things that you have to take into account when you are doing a risk assessment. Paragraph (f) says you have to take account of “the different ways in which the service is used, including functionalities or other features of the service that affect how much children use the service, (for example a feature that enables content to play automatically), and the impact of such use on the level of risk of harm that might be suffered by children”.  

Now, I accept that that does not include the word “addiction”, but it talks specifically about harm arising not from what the child sees but from how long the child is seeing it for, and the ways in which the service is designed, which keep the child on the service for a long time. That is exactly the kind of harm that I think you are describing.

If Ofcom chose to do so, it could say, “That gives me legislative authority to require of platforms that they address exactly that kind of risk in their risk assessment”. Again, it is worth keeping in mind the architecture of the Act; risk assessment is the foundation of almost everything. If you do the risk assessment properly and identify such risks, you are obligated to do something about them. So we have to get the risk assessment process right.

I am probably giving the impression that I am against amendments or improvements to this wording; I am not, and think there is scope for that in lots of ways. But we cannot suggest that this Act does not include any mention of addiction or design faults or design aspects—perhaps they are not faults; perhaps they are deliberate features—that keep you on the platform for longer than you might otherwise have been there. That is absolutely fundamental to this list, and what I have just read to you, for me at least, encompasses the concept of addiction. 

  1.        Viscount Colville of Culross: Beeban, I know that you feel really passionately about this, so I will come to you. You have already given Ofcom a good kicking over this particular issue, to say the least. Is there something more that needs to be done to move Ofcom into action? Do we need to define what addiction is and get that put into the Act? What needs to be done?

Baroness Kidron: I also brought the quote from the Government’s response to the amendment I put down: they said at the Dispatch Box that it did mean extended use, even when there was no harmful content. I have that exchange, and I will send it to the committee, in the interest of time. It is rather entertaining, because it is about marshmallows, but it makes the point that Jeremy has made in another way.

I would love for the committee to make it explicit. I do not think we need a definition of addiction—that is going down somebody else’s path—but I think extended use is a concept that we know within the framework and sector that we are talking about. You should get the “safety by design” bit right and absolutely specify that part of what you are looking at is the lost time and going places you did not really want to go—this is all extended use. There are a bunch of features, and I will send them to the committee as a list. The other way of saying it is that there are dark patterns that keep you there; that is covered by Section 11(6)(f), and Ofcom should have been more generous and ambitious in bringing it into the codes. If you do not deal with that, you will never even begin to deal with the content, because there is just more and more.

The Chair: Thank you. In the interest of time, we will have to move on.

  1.        Lord Kirkhope of Harrogate: I start by saying this: never underestimate the importance of adjectives. Sir Jeremy has referred to functionality as being a very important addition to the way we take these things forward. I used proportionality in the GDPR in Brussels. These are important issues, and I want short answers to this question. In your initial remarks, you all mentioned certain expectations that you had when you set out on this, particularly expectations of the appointed regulator and what it would do to interpret the legislation and implement it in a way that really satisfied you. In one or two sentences each, can you tell me this: in which areas where you had high expectations have you felt let down by the regulator and the implementation of this?

Damian Collins: As I said at the beginning, it is proactive scanning for content. My view is that Ofcom’s evidence to the committee is at total odds with what is written in the legislation and what we expected at the time, and it is fundamental to this working.

  1.        Lord Kirkhope of Harrogate: That is a good answer. Thank you. Can I have answers from everybody else, too, please?

Professor Lorna Woods: It is its approach to codes and evidence—evidence with the impact on codes.

Baroness Kidron: It is the way that it has cherry-picked the Act. It did not do different ages and stages, or risk and size—it only did size in its categorisation. It did not use its information-gathering powers as a tool to understand the evidence, as it thought it was disproportionate to use its own powers. The last area is enforcement, enforcement, enforcement.

  1.        Lord Kirkhope of Harrogate: Jeremy, do you have an adjective?

Sir Jeremy Wright: At the root, it is a minimalist interpretation of the Act, which I do not think is merited. That is it in a nutshell, for me. Ofcom has looked at this Act and interpreted it extraordinarily restrictively in my judgment—far more restrictively than the language allows for. As a result, it has retreated far within the boundaries of the legislation. It could go much further up to those boundaries and still remain perfectly within the scope of the Act.

The Chair: Thank you. Some of our questions overlap, so do not feel you have to answer if you do not have anything specific to say. We will need to keep our answers pretty short, just as you have been doing there—although we have had quite a richness of answers so far, so I thank you for that.

  1.        Baroness Fleet: Thank you for all the work you have all done on this really knotty issue. I wanted to come to Ofcom’s powers and how they should be strengthened. Because of the shortage of time, I would like to hear your views—briefly, if possible—on whether Ofcom really has the power to use the business disruption measures that have been referred to and what more could be done. We can see already that the fines are not really having any impact at all: they are trifling in scale and can be easily dismissed. Do you think business disruption measures can and should be enforced?

Sir Jeremy Wright: The answer is yes, they should be used when it is appropriate to use them. The truth is that we do not know yet about their effectiveness, because they have hardly been used, and so it is hard to tell. Whatever the figures for fines allocated so far, the maximum allowable is 10% of global turnover—that is a serious amount of money—and there are criminal penalties, including imprisonment, available in some circumstances as well. I think Ofcom has the sanctions that it needs.

If you are asking me whether there is a gap that needs filling in Ofcom’s powers, I would say that, probably, the bit of this Act that is least talked about and needs to be talked about more is the media literacy part of it—in other words, what do we do to equip our citizens to deal more adequately with the digital threats that they face? This is not just about offensive material online in social media platforms; it is also about mis- and dis-information, some of it deliberate and state sponsored. There is a real vacancy in this legislation for Ofcom having more power to do things a little more dynamically around filling in those gaps. If it is not to be done in this legislation, it needs to be done elsewhere. I am all in favour of Ofcom and others doing more and having more power to do more on the combating of mis- and dis-information, which rather got forgotten about in this general package of legislation.

  1.        Baroness Fleet: Professor Woods, can you speak specifically about whether the business disruption measures could now be used if it makes a case?

Professor Lorna Woods: There are a couple of points. Picking up on Jeremy’s point that they have not been used, we are in a space where there is some nervousness about the level of evidence that would be required to satisfy the court that either type of business disruption measure is justified. I do not know how we get out of that circular trap. The business disruption measures aimed at ancillary services are potentially more useful than access restrictions, because they do not have the freedom of expression costs that cutting off services would have, but I question whether having exactly the same threshold of evidence for the service restriction orders as for the access restriction is quite right. Ofcom has pointed out that it has the same evidence threshold to use the business disruption as an interim measure as for a final measure; you would think that, if it is an interim measure, it possibly will not have been in the position to have that evidence to use for interim measures. Again, there may be some usefulness in looking at thresholds.

Damian Collins: The short answer to your question is yes, but I can see why there is the question of the point at which that threshold is met where those measures can be introduced. You can understand why Ofcom may have some caution here: it is currently being judicially reviewed by Meta with regard to whether the fines and fees regime in the legislation is proportionate, so it probably anticipates other legal challenge when any business disruption measures are brought forward.

It begs one other question as well, which goes back to a recommendation that the Joint Committee made at the time about having a clearer route for civil redress through the courts for victims. That could be a bigger deterrent than some of the business disruption measures: as we see in America, it potentially comes with a much bigger bill. I really welcome the review that the Law Commission has launched now on consumer redress. A consumer redress regime that would allow class action litigation and litigation by victims of online safety against the platforms would be a very welcome addition. It would create more financial jeopardy for the companies themselves and recognises that Ofcom plays a significant role but cannot do it all by itself—just as, in competition law, people can bring a follow-on action on behalf of a CMA ruling or can bring their own independently. That would be a welcome addition here.

Baroness Kidron: Can I add just a very small thing about business disruption measures? One of the requirements that Ofcom has in the law is that once something is settled it no longer has a case against the company. It has been put into situations where it has chased after a company for years and then at the last minute it is geo-blocked and there is nothing it can do. There are some technical barriers.

I want to leave the committee with a couple of things. I absolutely agree about the routes of civil action, but the Law Commission is also looking at product liability and it would be very powerful to extend it to digital services and products. I really want to make the point that there is nowhere for an individual to go. I dealt with a family who were having an absolute crisis with a child and a chatbot. I called Ofcom—I will not say who—because the family had been told by the police that, because there was no perpetrator, they had no part to play. When I spoke to Ofcom, it said, “We do not have any purchase on this. We do systems and processes, so they can just put it into our horizon-scanning email”. That is not good enough in two ways. One is because, in this report it wrote last week, it says it may have some purchase over chatbots, but not all purchase so it is going to wait. Secondly, it is not okay for us as a society to have nowhere to go. I would also recommend an e-safety commissioner or someone who takes on individuals and is willing to fight on behalf of immediate danger.

Again, I do not want to take your time, but I have a few detailed proposals around enforcement, which I think is where we should seek to give Ofcom much more support. As we have been so critical about what it could do, we must also make sure it has the tools to do what we want it to do.

The Chair: If you could send those to the committee, that would be very helpful.

Baroness Kidron: I will do.

The Chair: That would be very helpful, thank you. I am just aware of the time. We are okay to carry on, but I am just checking with you, Baroness Kidron: are you okay to do so?

Baroness Kidron: I am fine.

  1.        Lord Knight of Weymouth: We have talked about how the world has changed since the Act was passed and how new technologies and new harms are being developed and used all the time. That raises the question of future-proofing the legislation, which has been touched on. The Government have responded by taking significant powers of regulation, including being able to change the primary legislation—Henry VIII powers, as we call them here. Do you think that is a sustainable way to create agility or do we need new primary legislation or something else? Do we need a permanent Joint Committee to help advise both Houses, as the Joint Committee suggested? Do we need some new regulatory arrangements so that in certain circumstances we can have regulation that is amendable by Parliament? I will start with Jeremy, and then I particularly want to hear from Beeban and, if we have time, from Lorna and Damian.

Sir Jeremy Wright: In your list, not all of those are alternatives. I am very much in favour of the Joint Committee model. It is really important that we can maintain scrutiny of what regulators do with our legislation. What we have all been discussing this afternoon is a piece of legislation which left an awful lot of latitude to the regulator, I think necessarily. I do not think that is a flaw in the system; I think that is what we had to do. But what do you do as Parliament if you hand over a great deal of power to a regulator and the regulator does not do what you anticipated it would do with it? What is Parliament’s comeback to that? The answer lies in constant and detailed scrutiny which, to be frank, the existing parliamentary structures find very difficult to provide. Both this committee and its equivalent in the House of Commons have many other things to do, and therefore Damian’s committee was right to propose a Joint Committee model which would look at digital regulation and be able to keep track of it as the circumstances change. It was right then; it is still right.

But that does not mean that you do not need to think about legislative improvement as well, because any such committee would not have legislative power in and of itself. It is not the case that we will never do another Act on this subject but, to be honest, I bear the scars of this one taking the better part of six years. I will accept that there were some changes of Prime Minister that delayed things, but the political class is going to be very nervous about undertaking a new piece of primary legislation. As we have been saying throughout, it is very hard, actually, to make a case for a new piece of primary legislation if you are not using fully the one you have got. We will have to do that first before the case for further primary legislation is required.

I am not in favour of Henry VIII powers as a matter of principle. But inevitably in this space, you have a choice—either Parliament retains the opportunity to keep up with the constant changes you described, or you need to devolve that authority to the regulator. It is neater for the regulator to do it, but with the additional safeguard that Parliament keeps a closer eye on what the regulator is doing than currently.

Baroness Kidron: I am strongly in favour of the Joint Committee idea. In both ends of Parliament, there is a deep knowledge about technology and how it works and how it could work better. By the very nature of government and our civil service, that depth of knowledge is not in government because they flow through. It would be such a harnessing of useful knowledge and oversight to do it that way.

Ofcom has come in for quite a bashing today from all of us, but there are other regulators, including the ICO, which have some place in this which have also done poorly. They did not always. I do not want to sit here and say regulation does not work, because the ICO under Elizabeth Denham worked brilliantly. We need to look in the long term towards a digital regulator that brings some of those things together. That bit of the AI Security Institute that just got canned—the societal bit—should be brought in as well, because we need that depth of understanding in a bigger regulator.

There are two dependencies. The first is: this is an emergency and now we should be doing everything we can do, immediately, swiftly, cleanly and clearly. Those bits that are political are on the Government to change the regulatory direction of Ofcom. They need a few powers and probably have enough Henry VIII powers to get them in.

This sector has brought US-style lobbying into our parliamentary system. I note for the record that both government and regulator are heavily lobbied on everything to do with this sector, not just child safety. While we are talking about child safety and online safety, the safety of United Kingdom citizens is being put at risk in relation to trade deals, relationships and investment deals. We have to really understand that that is happening in order to work out what it is that we are asking for.

Damian Collins: When we recommended the permanent Joint Committee when we were on the Joint Committee—a lot of the work had been done by the predecessor committee to this committee as well—it was to not just oversee the way Ofcom uses powers, but also to improve scrutiny of the co-ordination between different regulators. There was a discussion at the time about having an internet regulator, almost like the internet was a discrete thing rather than being part of everything. It is more sensible to look at how different regulators that work across the online world, be they the ICO or Ofcom, use their powers, and whether that is what Parliament intended. It is also recognising that, while we have departmental Select Committees, they are busy with lots of other important issues. The ability to build up a field of expertise and knowledge in this regulatory environment over time, and for there to be a permanently sitting Joint Committee, would have been of huge value. I congratulate this committee on doing this post-legislative scrutiny of the Act when no other attempt has been made to do that since it was passed.

The principal architecture of the Act—the duty of care based on the priority illegal harms—was actually the recommendation of the Joint Committee; it was not the way the Government intended to structure it. It was recommended with the future in mind, to make it easy for Parliament to add those offences in through secondary legislation. It was always envisaged, I think rightly, that future enforcement on new harms should be as a consequence of an Act of Parliament as well as a decision by the regulator. Parliament has to decide whether or not this is an offence or not, as Parliament has done over Baroness Owen’s amendment on intimate image sharing. We did not consider AI as a tool to fabricate intimate images; we were more concerned at the time with sharing ones that had already been captured. But we can create a new offence: Parliament can decide what it is, it can include it in the Act as a priority legal offence, and then Ofcom has the regulatory duty to enforce it. Imagine if we did not have that architecture and we were basically going to Ofcom and saying, “We think you should do something about this. We haven’t decided whether it’s illegal or not, but we think you should stretch powers you do not quite have in order to include this”. Would we want to delegate that responsibility? Even if we did, it would almost certainly be challenged in the courts by the platforms themselves.

The Chair: Thank you, I think that has been a full answer. Baroness Owen has a question on a slightly different topic.

  1.        Baroness Owen of Alderley Edge: I am interested in international comparisons so I want to touch slightly on the Meta settlement, which Damian briefly mentioned. We have seen that it limits notifications to children and there is an option to choose a feed not driven by algorithms, but it only applies to certain participating US states, not to the UK. I am interested in what this means for the Online Safety Act. Is it a failure of the Act that it is not as comprehensive as the protections that children will now have in America?

Baroness Kidron: We have already talked about having routes to the courts and how effective that has been. In fact I brought with me, but have not had time to read out, all the things that came out in disclosures that companies have said about their own services that make you understand that they know that they are harming. I read that list, and there is nothing in it that we have not been saying to Meta for more than a decade, and there is very little in the list that would not be covered if you fully used the age-appropriate design code and the Online Safety Act; between them, they cover almost everything in that list. The failure that we are talking about now is a failure of implementation and intent. We need to be seeing this enforced rather than supervised. The courts are somewhat more ruthless in saying this or that.

This raises an issue that I am glad you brought us around to, which is the question of fines. That fine looked enormous, and most newspapers talked only about the fine rather than changes to design—yet, when Meta settled, its share price went up so much that it almost paid for the fine overnight. The question of fines is not the issue. It is about business disruption, market conditionality, design requirements and saying, “You can’t serve kids unless you do those things”. That is the model. The interesting piece of the Meta case is that for the first time they did not look at speech—they were not frightened of all the things that were swirling around the Online Safety Bill in the beginning. What they did was say, “This is a product liability issue. You know you are hurting kids. Stop it”.

Damian Collins: The cases in America rightly got a lot of interest here, because you could see that the US Congress finds it impossible to regulate or legislate even over the areas, like child safety and combating fraud, where there would be the most consensus, but the courts have just cut straight through that. They have cut through First Amendment rights and the Section 230 liability shield and said, “This is a business design feature that has been designed by the company to make money for itself, and therefore it has liability over that”. So Baroness Kidron is right that the courts are more ruthless. That is why here, if we had civil litigation alongside a regulatory regime, the courts could also interpret what they thought the safety standards should be. They could use their discovery powers to get information that maybe Ofcom does not currently have.

That is an area that we have not talked about at all. There is an oblique reference in the Act to a skilled person’s report, which gives Ofcom the power to bring in outside experts to forensically examine the companies and seek information relevant to its investigation of whether they are meeting their safety duties, rather like a discovery process in court would do. I do not really see any evidence of that taking place, but it could do. It gives us the ability to publicly shame the companies for things that we know they are doing and not acting on and the knowledge that exists within the companies about that. That is what the whistleblowers have uncovered and what the courts have discovered, and Ofcom should use the powers that it has in the Act to try to do more of that as well. However, the cases in America highlight where having a robust civil litigation system, with class action making it possible for litigants to bring their claims, alongside the regulatory structure would be very beneficial.

The Chair: Beeban, you mentioned some things that you wish that you had been able to say. Do send anything like that on to us.

Baroness Kidron: Yes, I will send you my quote list.

  1.        Baroness Caine of Kentish Town: Thank you all for your pioneering work in this area and for the very effective evidence that you have given us today. It has been iterative, as you have noticed, and I have the final iterative question. Considering everything that you have just said about the implementation of the Act, I am going to give you one priority each. What would be the key recommendations that you think the committee should make to improve the effectiveness of the UK’s online safety regime? If you can, please break that down into what you think might need secondary legislation or other legislation, like the civil actions, or whether you think it is within the current framework.

Sir Jeremy Wright: I will say nothing that I have not already said. If there is a case to be made for additional powers delivered through legislation, then the way to make it is to use the powers you already have and then come back and ask for more. We are not yet at that point, so I remain of the view that we can do more with the Act that we have. The political realities of this mean that we are not going to see primary legislation any time soon, however much we may all want it and however many flaws we can see in this. The Act is what we are stuck with for the foreseeable future, so in my view we have to make the best of it, and this committee can do hugely good work in setting out the ways in which we can move further in the context of the legislation that we already have.

Professor Lorna Woods: I will cheat a little, if I may, and say that my one thing is a list of 10 things. I am quite prepared to send the list to the committee afterwards. It covers some of what we have talked about and a few more issues, all about tidying up the existing framework and, in a way, seeking to reassure Ofcom or to clarify the differences in interpretation between what perhaps the understanding in Parliament and civil society was and where Ofcom has got to.

I will highlight for the committee that the assumption in 2019, and maybe in 2023, was that we were dealing with a baseline that was going to go up but actually it has gone down. Some of that is new technology enabling new forms of harm, but some of it is companies withdrawing protections. We need to stop and say that actually 2023 was a baseline, and we cannot go backwards.

Damian Collins: There has been some criticism of Ofcom, but it has also been pretty honest with this committee about where it thinks the limitations of its powers lie. We disagree with that on the basis that we thought the Act gave it more power than it is using but, if there is a genuine concern that the powers do not work in quite the way that Parliament intended, how do we address that? How do we give Ofcom the cover it needs to do that? As I said earlier on, the question of proactive responsibilities—monitoring the responsibilities that the platforms should have—is central to the way that the Act should work, and that has to be fixed.

I have always thought that the skilled person’s report has a lot of scope. Are there questions about funding that work or what triggers an investigation? If that could be resolved, it would be helpful.

On the enforcement of the platforms’ terms of service, Ofcom clearly thinks that the platforms have to demonstrate that they act proportionately in the way that they seek to enforce them. Again, our sense was that they should go much further than that and make sure that they really do. So is that a gap in law that needs to be corrected or just something that is built in through the way the codes were developed? That will be fundamental to the question of whether you can enforce a children’s ban on social media. What do the companies have to do to demonstrate to the regulator how this can or cannot be done?

We technically have an under-13s’ or under-14s’ ban on social media, depending on the age of the app, but Ofcom does not really seem to have any ability to make the companies enforce those, so we have already failed in that regard. Before we go further, how can we be convinced that the regulator will have the power to enforce that?

  1.        Baroness Caine of Kentish Town: Beeban, could you encompass what we do with companies that are ideologically unwilling?

Baroness Kidron: Everything that we have said is what we would like the committee to recommend, but if I had to pick one thing then I would say that the outcome should be that Ofcom has to enforce against every risk in the risk register. Every risk in the risk register has to be mitigated, managed or forbidden, and we need to use Henry VIII powers, what is already in the Bill, government direction and willingness and, if necessary, new powers for Ofcom to do so.

The Chair: We could have probably gone on for another hour. I am sorry we did not have time to get to everything on your list, but do send us your list of 10 things, your quotes and anything else that you wanted to raise. We have quite a bit more evidence gathering to do, and then we are going to move on to Ofcom and the Minister. We will be gathering pace as we get towards the end of the year, and early next year we should be making our recommendations. Thank you; that has been very helpful.