Baroness Jones of Whitchurch:

Moved by Baroness Jones of Whitchurch

That this House do not insist on its disagreement with the Commons in their Amendment 32, on which the Commons have insisted for their Reason 32D, and do not insist on its Amendments 32B and 32C proposed to the words restored to the Bill by the Lords disagreement, to which the Commons have disagreed for the same Reason.

32D: Because it is not appropriate to require the Secretary of State, in preparing the DVS trust framework, to carry out an assessment of whether listed public authorities reliably ascertain sex data.

My Lords, I will also speak to Motions B and D. This first group is concerned with amendments relating to sex and gender in digital verification services, the data dictionary and scientific research. In relation to digital verification services and the data dictionary, I am grateful to the noble Viscount, Lord Camrose, for his continued engagement on the issue of sex data. Although we are not dealing with amendments in lieu today, I will take this opportunity to address some misunderstandings that I fear sit behind the concerns of noble Lords which were raised in previous debates.

This Bill does not create one digital identity app or system that lists attributes such as gender that those wanting to verify information about someone are required to accept. Instead, it creates a legislative structure of standards, governance and oversight for digital verification services. It is possible to create a reusable digital identity. However, when an organisation chooses to use a DVS, it will enter into a contract with that provider; that contract will specify which attributes the organisation needs to verify and how the DVS will do it. Reusable digital identities can therefore be reused only when an organisation accepts in writing that they meet its needs. If a reusable digital identity verified gender, it could not be used to verify biological sex in cases where that was needed instead.

Where a public authority is using a DVS, it remains the case that a contract will have to be entered into. This will again set out what types of information the DVS will be able to make checks against and for what purpose. This will ensure it is explicitly clear what information is being verified when a DVS relies on public authority data released through the information gateway. I hope this reassures noble Lords that gender data could not and would not be used to verify biological sex. Similarly, individuals would not be able to reuse a digital ID verifying gender to verify biological sex.

It is for these reasons that I have laid the Motions to agree with the elected House, which removed Lords Amendments 32B, 32C, 52B and 52C. I am grateful to the Opposition for accepting the assurances offered and not tabling a Motion to insist on the previous amendments.

In response to last week’s debate, I would like to respond to concerns raised by a few noble Lords around public data when sex and gender data appear in the same field. Existing legislation already requires those processing personal data to ensure that the data they process is accurate for the purpose for which it is being used. This means that personal data processed as part of a digital verification check must be appropriate for the specific requirements of that check.

The contracts I have mentioned are a way to ensure compliance with this principle. Any personal data passed through the information gateway to DVS providers is a new instance of data processing, and therefore the data accuracy principle is reapplied. That principle requires that the personal data must not be misleading, which is of particular relevance given that public authorities will be sharing data for verification purposes. As Minister Bryant set out in the other place, if the Government identify an instance where a public authority is sharing gender data in a way that is misleading as to the fact that it cannot be used to verify biological sex, they will of course respond appropriately. In light of these reassurances and noting the clearly expressed view of the other place on these issues, I hope noble Lords will agree with Motion A.

On scientific research and Amendment 43B, I am grateful to the noble Viscount, Lord Colville, for the time he has afforded the Government on this issue and for our productive meeting last week. I hope to reassure him and other noble Lords that there are, as we have argued throughout, sufficient protections against the potential misuse of the term “scientific research”. It is not the effect of the provisions to provide blanket approval of the reuse of personal data for AI training under the banner of scientific research.

The policy intention behind the clauses is not to enable the reuse of personal data for AI training unless it is for genuine scientific research, which is set out in the criteria in the ICO guidance. As part of its Bill implementation work, the ICO will prepare revised guidance around processing for research purposes. I expect this will cover information on compliance for data protection principles, including the fairness and purpose limitation principles. This will include the reasonable expectations of data subjects for AI model training when it constitutes genuine scientific research.

As with the previous topic, I have tabled Motion B to agree with the Commons on this issue. I am grateful to the noble Viscount for not tabling an amendment in lieu. On this basis, I hope noble Lords will also agree with Motion B and secure the continued success of the UK’s scientific research sector. I beg to move.

My Lords, when I was Opposition Chief Whip in Another place, I was never really sure whether it was my job to make sure that legislation was as good as possible for the good of the country or as bad as possible in order to make sure that the Government were not re-elected. With this Bill, we have done our best to make sure that this legislation is better but, I am afraid, without success.

The noble Lord, Lord Vallance, said that you could not end up with two different sources of digital verification that showed two different biological sexes. Will the noble Baroness the Minister confirm that, because of the muddle that has existed for years on this, you could have two documents that are different: one document saying one gender and another saying the other? This Bill is a missed opportunity, although I shall not seek to divide the House on Motion A.

My Lords, I thank the Minister for her introduction. In view of the remarks made a week ago by the Minister, the noble Lord, Lord Vallance, who referred to government datasets from the past 15 years which mixed up sex and gender as “accurate”—or perhaps “sort of accurate”, because the exchange in the report varied slightly—do the Government defend the accuracy of those datasets, even though they were, and continue to be, muddled because no one knew what “sex” meant? Are we expected to rely on the accuracy of data which mixed up sex and gender—that is, male and female—or do the Government mean that we cannot defend those data because they were only sort of accurate? I am not entirely clear what the Government are telling us about relying on historic data.

I am also concerned about what insight this gives into what the Government intend to regard as accurate from now on. I continue to think that the Government are on quite a sticky wicket in regard to data accuracy on sex and gender and their refusal to enshrine true sex accuracy in this Bill. We continue to have a bit of a fudge, which shakes confidence in their intentions. This is a huge missed opportunity, but I realise we are not having a further vote.

I shall ask just one question. Clause 29 allows for the Secretary of State to publish supplementary codes for DVS providers. Will the Government commit to publishing a supplementary code to ensure that DVS providers understand how to verify sex accurately and avoid what has been described by the Government Benches as the “muddle” of the last 15 years?

My Lords, I thank all noble Lords who have contributed to this important debate. I will first speak to the issues around accurate recording of sex data before coming on to talk about scientific research.

Throughout the passage of the Bill, we have been clear that digital verification services will be a significant driver of data reliability and productivity. They are absolutely dependent on accurate recording and rigorous management of data. We supported my noble friend Lord Lucas in his original amendments on Report, and we tabled our own amendments from the front bench for Lords consideration of Commons amendments last week.

I am grateful to the Minister for her engagement on this issue, and I know she has taken our concerns seriously. That said, we remain concerned about the accurate recording and management of sex data, especially in light of the recent judgment of the Supreme Court. The Government must continue to remain vigilant and to take steps to ensure datasets held by the Government and arm’s-length bodies are, and continue to be, accurate.

On the definition of scientific research, I should put on record my thanks to the noble Viscount, Lord Colville, who has led on this. We absolutely recognise the points he has been raising, and his amendments have been both inventive and constructive, trying to deliver a better definition of scientific research on the face of the Bill. I am pleased that Ministers have engaged with him on this problem.

Despite our remaining concerns on both those issues, we feel we have made our views clear to Ministers and, although they have chosen not to act—which, I am afraid, we believe they may come to regret—we must be responsible in our scrutiny. We will not, therefore, oppose the Government’s Motions today.

My Lords, despite the fact this is not being pushed to a vote—I respect that, and I concede that the Government have made some clarifications, and potentially concessions, along the way in this debate—I think that the issue is not yet resolved. I call on the Government to try to solve this problem now, rather than leave it open to more years of muddle, confusion and misinterpretation, and that can happen away from here. I have noticed that the Government are not averse to using the odd statutory instrument, to which I am usually opposed; in this instance, I urge them to use a statutory instrument to sort this out. I fear that, unless they do, it will undermine trust in the new system.

To clarify, we are looking to identify datasets that have muddled up sex and gender, such as data from HMPO and the DVLA, and those that have not, such as sex registered at birth. Because of that muddle, we cannot rely on those databases. Is that not the very point? We are trying at this point to provide clarity to DVS providers. By the way, this would not in any way result in outing individual transgender people when they are using the DVS system to prove their identity or other attributes, such as their age or whatever. We are trying to ensure that each database has some consistency. If a dataset allows some people to be recorded as the wrong sex, then the whole dataset is unreliable as a source of sex data.

It was very helpful that the Government clarified in the midst of this, for example, that an official document such as a passport, whatever is written on it, cannot be proof of a change of sex; it is simply a record of the way somebody wants to be identified and is no use as a reliable source of sex data. As I have said, there are other official documents such as the driving licence where that is not the case.

I would simply urge the Government, from their own point of view, so that we do not carry on having this muddle and confusion and so that this system becomes trusted, to make sure that they sort this out, even if they will not do so here and now.

My Lords, I thank the Minister for her engagement and for defining what genuine scientific research is. I hope very much that the AI companies, when using this extraordinary exemption, will listen to the Government, and that the Government will ensure that the policy is enforced. The trust of the people of this country would be lost if they felt that their data was being reused by AI companies simply for product enrichment and profit, rather than for genuine scientific research. I thank the noble Viscount, Lord Camrose, and the noble Lord, Lord Clement-Jones, for their parties’ support.

My Lords, I too thank the Minister for her introduction to the three Motions in this group.

On these Benches, we welcome the Supreme Court’s judgment on the meaning of “sex” in the Equality Act 2010. However, as Ministers have stressed—and we agree—it is paramount that we work through the implications of this judgment carefully and sensitively. As we have previously discussed, the EHRC is currently updating its statutory guidance.

Ministers have previously given assurances that they are engaged in appropriate and balanced work on data standards and data accuracy, and we accept those assurances. They have given a further assurance today about how the digital verification services framework will operate. We rely on those ministerial assurances. In summary, we believe that the previously proposed amendments were premature in the light of the EHRC guidance and that they risk undermining existing data standards work. On that basis, we support the Minister in her Motions A and D.

Turning to Motion B, the noble Viscount, Lord Colville, will not press his Amendment 43B at this stage, as he intends to accept the assurances given by Ministers. We have consistently supported the noble Viscount’s efforts to ensure that scientific research benefiting from the Bill’s provisions for data reuse is conducted according to appropriate ethical, legal and professional frameworks. The Government have given significant assurances in this area. We understand that their position is that the Bill does not alter the existing legal definition or threshold for what constitutes scientific research under UK GDPR. The Bill does not grant any new or expanded permissions for the reuse of data for scientific research purposes, and, specifically, it does not provide blanket approval for using personal data for training AI models under the guise of scientific research. The use of personal data for scientific research remains subject to the comprehensive safeguards of UK GDPR, including the requirement for a lawful basis, the adherence to data protection principles and the application of the reasonableness test, which requires an objective assessment.

The collection of assurances given during several stages of the Bill provides reassurance against the risk that commercial activities, such as training AI models purely for private gain, could improperly benefit from exemptions intended for genuine scientific research serving the public good. I very much hope that the Minister can reaffirm these specific points and repeat those assurances.

My Lords, I thank noble Lords for their contributions. I reassure your Lordships’ House that the Government are progressing workstreams focused on the accuracy and reliability of sex data in public authority datasets in a holistic and measured manner, as I have described in previous debates. We welcome the Supreme Court ruling, and are now working hard to consider those findings and the upcoming guidance from the equalities regulator, which will help.

I reiterate that the trust framework requires DVS providers to comply with data protection legislation, including the data accuracy principle, where they use and share personal data. That includes the creation of reusable digital identities, as well as one-off checks. If they fail to comply with these requirements, they could lose their certification. This means that the sex information listed on a passport—which, as we all know, could be a combination of biological sex, legal sex under the Gender Recognition Act and gender identity—cannot be used to verify biological sex.

The noble Lord, Lord Arbuthnot, asked whether a person can have different genders appearing on different documents. Yes, you could have both genders appearing on different documents, but they could not be used to prove biological sex.

I should say to noble Lords that there is a requirement for all this information to be recreated, reused and rechecked each time. In response to noble Lords who asked about historic data, the data will be renewed and checked under the new information that is now available.

In the Majority of cases where DVS are used, there will not be a need to verify biological sex, as we have noted before, because many DVS requirements do not ask that question. Data sharing under the power created in Clause 45 will involve new processing of data, which must be in compliance with the data accuracy principle: that is, it must be accurate for the purpose for which the information will be used. Of particular relevance, given that public authorities will be sharing data for verification purposes, is the fact that data accuracy principles require that the personal data must not be misleading.

With regard to the question from the noble Baroness, Lady Ludford, about supplementary codes of practice, I can confirm that the trust framework already includes requirements on data accuracy for DVS providers. That framework will, of course, be updated from time to time.

On scientific research, let me repeat my thanks to the noble Viscount, Lord Colville, for his contribution on this issue. I am glad that he was reassured by my remarks that we have been able to come to an agreeable resolution. I very much concur with the comments of the noble Lord Clement-Jones, that there has to be an ethical basis to those standards, and that point is absolutely well made.

On that basis, I hope I have reassured noble Lords. I commend the Motion to the House.

My Lords, I wonder, with regard to sexuality, whether the Minister has considered those children who are, unfortunately, born with perhaps an ovary and a testis, or with genitalia which are difficult to identify. How do those become categorised under this regulation?

The second thing is that the definition of science proposed in the Bill is not science; it is technology, and there is a big difference, as I explained in the last speech. Science involves knowledge, and we do not know that knowledge until we have the knowledge. We cannot act on that knowledge until we know what the knowledge is. That is hugely important and, as the noble Lord, Lord Vallance, who is not now in his place, said, this has the risk of holding up research which is really necessary.

Before I close, I mention just one example of this to the noble Lord, Lord Clement-Jones. He made a rather derogatory point about my comment on infection. I did not point out to him that, when I was seven, my father came home with a mild bronchial infection, which went on to be pneumonia. After six months with various inadequate antibiotics—because they did not understand the dosage—penicillin did not work and he died of a brain abscess when I was just eight. That is an example of where research is needed continuously, even when we do not know what we are doing. It is very important to understand that. This Bill and its wording do not fully define science satisfactorily, certainly to scientists.

My Lords, the first thing I would say about categorisation, as I hope I have stressed all along, is that data verification services will be required to provide accurate information. Normally, biological sex is not one of the things that most people need for their identity most of the time, but there are provisions under DVS for categorising to take account of those variations. I talked about biological sex, legal sex under the Gender Recognition Act and gender identity, for example. I hope that my noble friend has taken on board that point.

We have a fantastic scientific research community in this country, and it is our intention that it will thrive and grow. We absolutely intend to provide the proper underpinning of that, so that the scientific community does not feel that it is being undermined. I can reassure my noble friend that the provision in this Bill does not undermine the scientific research community, and it can remain confident that it will be protected going forward.

Motion A agreed.