Online Safety Bill Debate
Full Debate: Read Full DebateDepartment: Department for Digital, Culture, Media & Sport
Baroness Kidron
Main Page: Baroness Kidron (Crossbench - Life peer)Department Debates - View all Baroness Kidron's debates with the Department for Digital, Culture, Media & Sport
Online Safety Bill
Baroness Kidron ExcerptsThursday 22nd June 2023
(10 months, 2 weeks ago)
Lords ChamberRead Full debate Online Safety Act 2023 View all Online Safety Act 2023 Debates Read Hansard Text Watch Debate Read Debate Ministerial Extracts Amendment Paper: HL Bill 87(Rev)-X Tenth marshalled list for Committee - (20 Jun 2023)
Baroness Kidron (CB)
My Lords, I too want to support this group of amendments, particularly Amendment 234, and will make just a couple of brief points.
First, one of the important qualities of the online safety regime is transparency, and this really speaks to that point. It is beyond clear that we are going to need all hands on deck, and again, this speaks to that need. I passionately agree with the noble Baroness, Lady Fox, on this issue and ask, when does an independent researcher stop being independent? I have met quite a lot on my journey who suddenly find ways of contributing to the digital world other than their independent research. However, the route described here offers all the opportunities to put those balancing pieces in place.
Secondly, I am very much aware of the fear of the academics in our universities. I know that a number of them wrote to the Secretary of State last week saying that they were concerned that they would be left behind their colleagues in Europe. We do not want to put up barriers for academics in the UK. We want the UK to be at the forefront of governance of the digital world, this amendment speaks to that, and I see no reason for the Government to reject it.
Finally, I want to emphasise the importance of research. Revealing Reality did research for 5Rights called Pathways, in which it built avatars for real children and revealed the recommendation loops in action. We could see how children were being offered self-harm, suicide, extreme diets and livestream porn within moments of them arriving online. Frances Haugen has already been mentioned. She categorically proved what we have been asserting for years, namely that Instagram impacts negatively on teenage girls. As we put this regime in place, it is not adequate to rely on organisations that are willing to work in the grey areas of legality to get their research or on whistleblowers—on individual acts of courage—to make the world aware.
One of the conversations I remember happened nearly five years ago, when the then Secretary of State asked me what the most important thing about the Bill was. I said, “To bring a radical idea of transparency to the sector”. This amendment goes some way to doing just that.
Baroness Harding of Winscombe (Con)
My Lords, I, too, support Amendments 233 and 234, and Amendment 233A, from the noble Lord, Lord Allan. As the noble Baroness, Lady Kidron, said, it has been made clear in the past 10 days of Committee that there is a role for every part of society to play to make sure that we see the benefits of the digital world but also mitigate the potential harms. The role that researchers and academics can play in helping us understand how the digital world operates is critical—and that is going to get ever more so as we enter a world of large language models and AI. Access to data in order to understand how digital systems and processes work will become even more important—next week, not just in 10 years’ time.
My noble friend Lord Bethell quite rightly pointed out the parallels with other regulators, such as the MHRA and the Bank of England. A number of people are now comparing the way in which the MHRA and other medical regulators regulate the development of drugs with how we ought to think about the emergence of regulation for AI. This is a very good read-across: we need to set the rules of the road for researchers and ensure, as the noble Baroness, Lady Kidron, said—nailing it, as usual—that we have the most transparent system possible, enabling people to conduct their research in the light, not in the grey zone.
Lord Parkinson of Whitley Bay (Con)
It is guidance rather than direction, but it will be done openly and transparently. Users will be able to see the guidance which Ofcom has issued, to see whether companies have responded to it as they see fit and, through the rest of the framework of the Bill, be empowered to make their decisions about their experiences online. This being done openly and transparently, and informed by Ofcom’s research, will mean that everyone is better informed.
We are sympathetic to the amendment. It is complex, and this has been a useful debate—
Baroness Kidron (CB)
I wonder whether the Minister has an answer to the academic community, who now see their European colleagues getting ahead through being able to access data through other legislation in other parts of the world. Also, we have a lot of faith in Ofcom, but it seems a mistake to let it be the only arbiter of what needs to be seen.
Lord Parkinson of Whitley Bay (Con)
We are very aware that we are not the only jurisdiction looking at the important issues the Bill addresses. The Government and, I am sure, academic researchers will observe the implementation of the European Union’s Digital Services Act with interest, including the provisions about researchers’ access. We will carefully consider any implications of our own online safety regime. As noble Lords know, the Secretary of State will be required to undertake a review of the framework between two and five years after the Bill comes into force. We expect that to include an assessment of how the Bill’s existing transparency provisions facilitate researcher access.
Lord Clement-Jones (LD)
My Lords, I congratulate the noble Baroness on having elucidated this arcane set of amendments. Unfortunately, though, it makes me deeply suspicious when I see what the amendments seem to do. I am not entirely clear about whether we are returning to some kind of merits-based appeal. If so, since the main litigators are going to be the social media companies, it will operate for their benefit to reopen every single thing that they possibly can on the basis of the original evidence that was taken into account by Ofcom, as opposed to doing it on a JR basis. It makes me feel quite uncomfortable if it is for their benefit, because I suspect it is not going to be for the ordinary user who has been disadvantaged by a social media company. I hope our brand spanking new independent complaints system—which the Minister will no doubt assure us is well on the way—will deal with that, but this strikes me as going a little too far.
Baroness Kidron (CB)
My Lords, I enter the fray with some trepidation. In a briefing, Carnegie, which we all love and respect, and which has been fantastic in the background in Committee days, shared some concerns. As I interpret its concerns, when Ofcom was created in 2003 its decisions could be appealed on their merits, as the noble Lord has just suggested, to the Competition Appeal Tribunal, and I believe that this was seen as a balancing measure against an untested regime. What followed was that the broad basis on which appeal was allowed led to Ofcom defending 10 appeals per year, which really frustrated its ability as a regulator to take timely decisions. It turned out that the appeals against Ofcom made up more than 80% of the workload of the Competition Appeal Tribunal, whose work was supposed to cover a whole gamut of matters. When there was a consultation in the fringes of the DEA, it was decided to restrict appeal to judicial review and appeal on process. I just want to make sure that we are not opening up a huge and unnecessary delaying tactic.
Viscount Camrose (Con)
I thank all those who have spoken, and I very much appreciate the spirit in which the amendments were tabled. They propose changes to the standard of appeal, the standing to appeal and the appeals process itself. The Government are concerned that enabling a review of the full merits of cases, as proposed by Amendments 243 and 245, could prove burdensome for the courts and the regulator, since a full-merits approach, as we have been hearing, has been used by regulated services in other regulatory regimes to delay intervention, undermining the effectiveness of the enforcement process. With deep-pocketed services in scope, allowing for a full-merits review could incentivise speculative appeals, both undermining the integrity of the system and slowing the regulatory process.
While the Government are fully committed to making sure that the regulator is properly held to account, we feel that there is not a compelling case for replacing the decisions of an expert and well-resourced regulator with those of a tribunal. Ofcom will be better placed to undertake the complex analysis, including technical analysis, that informs regulatory decisions.
Amendment 245 would also limit standing and leave to appeal only to providers and those determined eligible entities to make super-complaints under Clause 150. This would significantly narrow the eligibility requirements for appeals. For appeals against Ofcom notices we assess that the broader, well-established standard in civil law of sufficient interest is more appropriate. Super-complaints fulfil a very different function from appeals. Unlike appeals, which will allow regulated services to challenge decisions of the regulator, super-complaints will allow organisations to advocate for users, including vulnerable groups and children, to ensure that systemic issues affecting UK users are brought to Ofcom’s attention. Given the entirely distinct purposes of these functions, it would be inappropriate to impose the eligibility requirements for super-complaints on the appeals system.
I am also concerned about the further proposal in Amendment 245 to allow the tribunal to replace Ofcom’s decision with its own. Currently, the Upper Tribunal is able to dismiss an appeal or quash Ofcom’s decision. Quashed decisions must be remitted to Ofcom for reconsideration, and the tribunal may give directions that it considers appropriate. Amendment 245 proposes instead allowing the Upper Tribunal to
“impose or revoke, or vary the amount of, a penalty … give such directions or take such other steps as OFCOM could itself have given or taken, or … make any other decision which OFCOM could itself have made”.
The concern is that this risks undermining Ofcom’s independence and discretion in applying its powers and issuing sanctions, and in challenging the regulator’s credibility and authority. It may also further incentivise well-resourced providers to appeal opportunistically, with a view to securing a more favourable outcome at a tribunal.
On that basis, I fear that the amendments tabled by the noble Lord would compromise the fundamental features of the current appeals provisions, without any significant benefits, and risk introducing a range of inadvertent consequences. We are confident that the Upper Tribunal’s judicial review process, currently set out in the Bill, provides a proportionate, effective means of appeal that avoids unnecessary expense and delays, while ensuring that the regulator’s decisions can be thoroughly scrutinised. It is for these reasons that I hope the noble Baroness will withdraw the amendment.
Lord Knight of Weymouth (Lab)
The fact that I labelled it as being AI-generated helped your Lordships to understand, and the transparency eases the debate. I beg to move.
Baroness Kidron (CB)
My Lords, I thank the noble Lord, Lord Knight, for laying out the amendment and recognise that there was a very thoughtful debate on the subject of machine-generated content on Amendment 125 in my name on a previous day of Committee.
I appreciate that the concept of labelling or watermarking machine-generated material is central to recent EU legislation, but I am equally aware that there is more than one school of thought on the efficacy of that approach among AI experts. On the one hand, as the noble Lord, Lord Knight, beautifully set out—with the help of his artificial friend—there are those who believe that visibly marking the division of real and altered material is a clue for the public to look more carefully at what they are seeing and that labelling it might provide an opportunity for both creators and digital companies to give greater weight to “human-created material”. For example, it could be that the new BBC Verify brand is given greater validity by the public, or that Google’s search results promote it above material labelled as machine-generated as a more authentic source. There are others who feel that the scale of machine-generated material will be so vast that this labelling will be impossible or that labelling will downgrade the value of very important machine-generated material in the public imagination, when in the very near future it is likely that most human activity will be a blend of generated material and human interaction.
I spent the first part of this week locked in a room with others at the Institute for Ethics in AI in Oxford debating some of these issues. While this is a very live discussion, one thing is clear: if we are to learn from history, we must act now before all is certain, and we should act with pragmatism and a level of humility. It may be that either or both sets of experts are correct.
Industry has clearly indicated that there is an AI arms race, and many companies are launching services that they do not understand the implications of. This is not my view but one told to me by a company leader, who said that the speed of distribution was so great that the testing was confined to whether deploying large language models crashed the platforms; there was no testing for safety.
The noble Lord, Lord Stevenson, says in his explanatory statement that this is a probing amendment. I therefore ask the Minister whether we might meet before Report and look once again at the gaps that might be covered by some combination of Amendment 125 and the amendment in front of us, to make certain that the Bill adequately reflects the concerns raised by the enforcement community and reflects the advice of those who best understand the latest iterations of the digital world.
The Communications Act 2003 made a horrible mistake in not incorporating digital within it; let us not do the same here. Adding explicit safety duties to AI and machine learning would not slow down innovation but would ensure that innovation is not short-sighted and dangerous for humanity. It is a small amendment for what may turn out to be an unimaginably important purpose.
Lord Allan of Hallam (LD)
My Lords, it is a pleasure to follow the noble Baroness, Lady Kidron. I will try to keep my remarks brief.
It is extremely helpful that we have the opportunity to talk about this labelling question. I see it more as a kind of aperitif for our later discussion of AI regulation writ large. Given that it is literally aperitif hour, I shall just offer a small snifter as to why I think there may be some challenges around labelling—again, perhaps that is not a surprise to the noble Baroness.
When we make rules, as a general matter we tend to assume that people are going to read them and respond in a rationalist, conformist way. In reality, particularly in the internet space, we often see that there is a mixed environment and there will be three groups. There are the people who will look at the rules and respond in that rational way to them; a large group of people will just ignore them—they will simply be unaware and not at all focused on the rules; and another group will look for opportunities to subvert them and use them to their own advantage. I want to comment particularly on that last group by reference to cutlery and call centres, two historic examples of where rules have been subverted.
On the cutlery example, I am a Sheffielder, and “Made in Sheffield” used to mean that you had made the entire knife in Sheffield. Then we had this long period when we went from knives being made in Sheffield to bringing them to Sheffield and silver-plating them, to eventually just sharpening them and putting them in boxes. That is relevant in the context of AI. Increasingly, if there is an advantage to be gained by appearing to be human, people will look at what kind of finishing you need, so: “The content may have been generated by AI but the button to post it was pushed by a human, therefore we do not think it is AI because we looked at it and posted it”. On the speech of the noble Lord, Lord Knight, does the fact that my noble friend intervened on him and the noble Lord had to use some of his own words now mean that his speech in Hansard would not have to be labelled “AI-generated” because we have now departed from it? Therefore, there is that question of individuals who will want something to appear human-made even if it was largely AI-generated, and whether they will find the “Made in Sheffield” way of bypassing it.
Interestingly, we may see the phenomenon flipping the other way, and this is where my call centres come in. If people go to a popular search engine and type in “SpinVox”, they will see the story of a tech company that promised to transcribe voicemails into written text. This was a wonderful use of technology, and it was valued on the basis that it had developed that fantastic technology. However, it turned out—or at least there were claims, which I can repeat here under privilege—that it was using call centres in low-cost, low-wage environments to type those messages out. Therefore, again, we may see, curiously, some people seeing an advantage to presenting content as AI-generated when it is actually made by humans. That is just to flag that up—as I say, it is a much bigger debate that we are going to have. It is really important that we are having it, and labelling has a role to play. However, as we think about it, I urge that we remember those communities of people who will look at whatever rules we come up with and say, “Aha! Where can I get advantage?”, either by claiming that something is human when it is generated by AI or claiming that it is generated by AI if it suits them when it was actually produced by humans.
Baroness Kidron
“age estimation” includes reference to an age range or an age expressed in years;“age verification” means the exact age of a person in years, months, and days or an established date of birth;”Member’s explanatory statement
This amendment defines the meaning of age assurance in the Bill to recognise it includes any test of age including but not limited to verification. Age verification means the exact age of a person in years, months, and days or a date of birth. Age estimation may refer to an age range or an age expressed in years. This is a definition of terms only: the intention is that Ofcom will produce guidance of what level of assurance is required in different settings.
Baroness Kidron (CB)
My Lords, we already had a long debate on this subject earlier in Committee. In the interim, many noble Lords associated with these amendments have had conversations with the Government, which I hope will bear some fruit before Report. Today, I want to reiterate a few points that I hope are clarifying to the Committee and the department. In the interests of everyone’s evening plans, the noble Lord, Lord Bethell, and the noble Baroness, Lady Harding, wish to associate themselves with these remarks so that they represent us in our entirety.
For many years, we thought age verification was a gold standard, primarily because it involved a specific government-issued piece of information such as a passport. By the same token, we thought age estimation was a lesser beast, given that it is an estimate by its very nature and that the sector primarily relied on self-declarations with very few checks and balances. In recent years, many approaches to age checking have flourished. Some companies provide age assurance tokens based on facial recognition; others use multiple signals of behaviour, friendship group, parental controls and how you move your body in gameplay; and, only yesterday, I saw the very impressive on-device privacy-preserving age-verification system that Apple rolled out in the US two weeks ago. All of these approaches, used individually and cumulatively, have a place in the age-checking ecosystem, and all will become more seamless over time. But we must ensure that, when they are used, they are adequate for the task they are performing and are quality controlled so that they do not share information about a child, are secure and are effective.
That is why, at the heart of the package of measures put forward in my name and that of the noble Lords, Lord Stevenson and Lord Bethell, and the right reverend Prelate the Bishop of Oxford, are two concepts. First, the method of measuring age should be tech neutral so that all roads can be used. Secondly, there must be robust mechanism of measurement of effectiveness so that only effective systems can be used in high-risk situations, particularly those of primary priority harms such as self-harm and pornography, and that such a measurement will be determined by Ofcom, not industry.
From my work over the last decade and from recent discussion with industry, I am certain that any regime of age assurance must be measurable and hold to certain principles. We cannot create a situation where children’s data is loosely held and liberally shared; we cannot have a system that discriminates against, or does not have automatic appeal mechanisms for, children of colour or those who are 17 or 19, who are at most likelihood of error. Systems should aim to be interoperable and private, not leave traces as children go from one service to another.
Each of the principles of our age-verification package set out in the schedule are of crucial importance. I hope that the Government will see the sense in that because, without them, this age checking will not be trusted. Equally, I urge the Committee to embrace the duality of age verification and estimation that the Government have put forward, because, if a child uses an older sibling’s form of verification and a company understands through the child’s behaviour that they are indeed a child, then we do not want to set up a perverse situation in which the verification is considered of a higher order and they cannot take action based on estimation; ditto, if estimation in gameplay is more accurate than tokens that verify whether someone is over or under 18, it may well be that estimation gives greater assurance that the company will treat the child according to their age.
I hope and believe that, in his response, the Minister will confirm that definitions of age assurance and age estimation will be on the face of the Bill. I also urge him to make a generous promise to accept the full gamut of our concerns about age checking and bring forward amendments in his name on Report that reflect them in full. I beg to move.
Lord Clement-Jones (LD)
My Lords, I associate these Benches with the introduction by the noble Baroness, Lady Kidron, support her amendments and, likewise, hope that they form part of the package that is trundling on its way towards us.
Lord Parkinson of Whitley Bay (Con)
I am very grateful to the noble Baroness for her amendment, which is a useful opportunity for us to state publicly and share with the Committee the progress we have been making in our helpful discussions on these issues in relation to these amendments. I am very grateful to her and to my noble friends Lord Bethell and Lady Harding for speaking as one on this, including, as is well illustrated, in this short debate this evening.
As the noble Baroness knows, discussions continue on the precise wording of these definitions. I share her optimism that we will be able to reach agreement on a suitable way forward, and I look forward to working with her, my noble friends and others as we do so.
The Bill already includes a definition of age assurance in Clause 207, which is
“measures designed to estimate or verify the age or age-range of users of a service”.
As we look at these issues, we want to avoid using words such as “checking”, which suggests that providers need to take a proactive approach to checking age, as that may inadvertently preclude the use of technologies which determine age through other means, such as profiling. It is also important that any definition of age assurance does not restrict the current and future use of innovative and accurate technologies. I agree that it is important that there should be robust definitions for terms which are not currently defined in the Bill, such as age verification, and recommit to the discussions we continue to have on what terms need to be defined and the best way to define them.
This has been a very helpful short debate with which to end our deliberations in Committee. I am very grateful to noble Lords for all the points that have been raised over the past 10 days, and I am very glad to be ending in this collaborative spirit. There is much for us still to do, and even more for the Office of the Parliamentary Counsel to do, before we return on Report, and I am grateful to it and to the officials working on the Bill. I urge the noble Baroness to withdraw her amendment.