Data Policy Digest

Jul 17, 2023

Jul 17, 2023

Welcome to the fourth edition of Connected By Data’s Data Policy Digest, your one stop shop for all things data and AI policy. You’ll be pleased to know it’s shorter than our previous epic - but there’s still been plenty going on.

If there’s something we’ve missed, something you’re up to that you’d like us to include next time or you have any thoughts on how useful the Digest is or could be, please get in touch via jonathan@connectedbydata.org. We’re on We’re on Twitter @ConnectedByData and @DataReform. You can also catch up on Digest #1, Digest #2 and Digest #3.

Data policy developments

Deeply DPDIB

via GIPHY

We’re not expecting the Data Protection and Digital Information (No 2) Bill to return to parliament before summer recess - we’ll probably meet it again either before or after party conference season in the autumn. It’s very likely to be ‘carried over’ to the next parliamentary session - it won’t finish all the bill stages before this parliamentary session ends (at some point in the autumn), so it’ll continue into the next one.

There are a few other DPDIB updates, though. The Open Rights Group and Big Brother Watch are among the UK-based organisations signing a letter, with various international civil society groups, warning the European Commission of the threat DPDIB poses to data adequacy. The Register has a write-up, and Politico’s Morning Tech newsletter (behind a paywall) reports some ‘jitters in Brussels’ about whether the UK’s data adequacy deal would survive DPDIB. It comes as the UK has been granted associate status with the Global Cross Border Privacy Rules (CBPR) Forum (we touched on that in our first Digest, with a thread from Javier Ruiz), and the EU adopted a data adequacy decision with the US.

And on the subject of Threads… Meta has delayed the launch of its new Twitter rival in the EU (it’s already here in the UK) owing to uncertainty over how it uses personal data. Google, meanwhile, is about to start rolling out its Bard chatbot in Europe after resolving regulatory concerns.

And happy 39th birthday to the original UK Data Protection Act!

Bills, bills, bills

Online Safety Bill Both DPDIB and the Digital Markets, Competition and Consumers Bill (see below) may get carried over; the Online Safety Bill has to complete its legislative passage this session, or else it falls. That said, the government is trying to extend the time it has from next week to 31 October, which implies the next session will start after that date. The Lords continue to amend the Bill, with the government’s own amendments (around protecting children, age verification, coroners and bereaved parents being able to access data and Ofcom research powers towards app stores) and one from Baroness Kidron (‘Algorithms that lead boys to Andrew Tate content targeted’, reports the BBC) getting much of the attention. As ever, Carnegie UK’s Maeve Walsh has much more on the second chamber shenanigans.

Ofcom is doing the right things to prepare but won’t be ready to fully roll out the new regulatory regime until 2025, according to a new National Audit Office report. And Ofcom has just kicked off a consultation on how to categorise regulated services (social media, search engines and certain other websites) under the OSB.

And in other OSB news… building on the OSB, DSIT has announced a review of online pornography … security and privacy researchers wrote a joint letter concerned about surveillance technologies in the Bill, as Apple joined others expressing encryption concerns … Wikipedia founder, Jimmy Wales, remained unimpressed, arguing that the bill will make the internet less safe… Damian Collins MP and Signal’s Meredith Whittaker debated encryption and surveillance on Channel 4 News … and researchers are concerned about their access to data from Threads.

Digital Markets, Competition and Consumers Bill The DMCCB has been making its way through Commons committee stage. The Centre for Policy Studies published a report worrying that the bill gives the Competition and Markets Authority ‘new, extensive and unchecked’ powers, becoming an ‘unregulated regulator’ (also on the right, the Institute for Economic Affairs wrote about competition in digital markets). Producers of news were more positive about the possibility that tech giants could be forced to pay for news they are producing. There’s also been a big digital competition story this week: Microsoft’s intended purchase of games giant Activision Blizzard. Regulators in the EU and US, as well as the UK, all originally had concerns about lessening competition in cloud gaming services; the EU then declared itself satisfied by Microsoft’s concessions; the US Federal Trade Commission remains opposed to a merger - it attempted to block the purchase, only for a judge to rule against the FTC, thought the FTC has today moved to appeal; and it remains to be seen how the UK’s CMA will react, having previously found against the acquisition. It is apparently taking the next six weeks to review its position - just as Microsoft and Sony reach a deal to keep ‘Call of Duty’ running on PlayStation.

And if that isn’t enough bills for you… The Culture, Media and Sport select committee has started pre-legislative scrutiny of the Draft Media Bill, which focuses on broadcasting and new technology. They heard from minister John Whittingdale last week.

AI got ‘rithm

In our last Digest, we linked to responses to the government’s AI regulation white paper consultation from Big Brother Watch, the TUC and the West Midlands Police Data Ethics Committee; three tests and a roundtable (the lesser-known Richard Curtis comedy) from the Ada Lovelace Institute; and of course our own submission, not forgetting an alternative AI white paper from the Public Law Project. Since then, Which have published their response; Ada have put together an event with AWO next week to launch a longer response; and the Equality and Human Rights Commission described safeguards as ‘inadequate’. Get in touch if your organisation has published a response.

The UK civil service has published its guidance to civil servants on how they could (and shouldn’t) use generative AI in their work (with summaries from the Central Digital and Data Office and Civil Service World also available). I’d nearly forgotten there was also guidance - with case studies - produced back in 2019. Oxford Insights have also been pondering the government use of LLMs, while Congress, the Russell Group universities and the Australian Government have also been thinking about how to use generative tools, and Wired has a piece on how to protect your privacy while using them.

As everyone tries to get to grips with generative AI… the Columbia Journalism Review has published a piece on how to report better on AI, as an AI-written story apparently caused chaos at tech news site Gizmodo; comedian Sarah Silverman is suing OpenAI for copyright infringement, as actors join screenwriters in striking, other creators have started rebelling and director Christopher Nolan has shared his thoughts on AI; and Creative Commons raised some eyebrows with its take on sharing creative assets for generative AI, as games platform Steam clamped down on AI-generated components.

What’s happening with the UK government’s promised summit on AI safety? The New Statesman thinks prime minister Rishi Sunak is lowering expectations, following his appearance in front of parliament’s Liaison Committee, while AI minister told the Archbishop of York - who was (rightly) concerned about a lack of civil society representation - that those congregating at the summit would constitute a ‘broad church’. Jack Stilgoe argues the autumn AI summit gives the UK a chance to lead on responsible AI - if it ensures all perspectives, not just entrepreneurs and scientists, are heard - while the Centre for the Governance of AI thinks the Summit should aim to produce shared commitments and consensus statements from states, plant the seeds for new international institutions, highlight UK AI policy initiatives, secure commitments from AI labs, increase awareness and understanding of AI risks and governance options, and commit participants to annual AI safety summits and further discussions.

Summit’s already happened in Geneva, where AI for Good generated a lot of headlines, from historian Yuval Noah Harari calling for firms creating fake humans to face prison and AI (though not education) expert Stuart Russell opining on the impact AI will have on education, to a press conference with robots. No, really. And on 18 July, the UK will lead a UN Security Council discussion on AI.

You’ll be aware (from our last edition) of various other UK AI announcements. If you want to know more about what the new chair of the Foundation Model Taskforce thinks about AI, then Adam recommends this podcast Ian Hogarth did with Novara Media just before his appointment. He’s also just given an interview to the BBC. And Responsible AI UK have just launched their first funding call - they’re looking ‘to support interdisciplinary translational research and knowledge exchange on responsible artificial intelligence (RAI) to ensure that AI technologies are designed, deployed and used responsibly within societies’. Deadline is 7 September.

Elsewhere in UK government, the Committee for Standards in Public Life - an independent committee that advises the Prime Minister on ethical conduct in public life - published a report on what AI might mean for standards in public life back in early 2020. (One suspects they’ve been pretty busy recently.) They’ve just written to government departments and public bodies, and to regulators asking for updates on how they’re adapting to the challenges posed by AI. The Electoral Commission is worried we’re running out of time to keep up with AI. And DWP says it’s using AI to catch people involved in benefit fraud. One welfare expert is sceptical. As are the National Audit Office, among others.

The latest Global AI Index from Tortoise sees the UK slip from third to fourth - largely because of a quickly-climbing Singapore. There’s lots of interesting insight in the detail. The World Economic Forum notes countries everywhere are under pressure to revisit their AI strategies.

And in other AI news… a couple of academic papers were doing the rounds on what’s left of Twitter, on international institutions for advanced AI and AI regulation … OpenAI are opening an office in London and have been talking about superalignment … in the US, the FTC has generative AI competition concerns, the Washington Post has been counting the financial cost of generative AI and the Weather Network has been counting the environmental cost … Sky News ran a special AI Future programme … the BBC’s Visual Journalism Team have an introduction to AI … China has issued new rules on generative AI … and not content with challenging one another in the ring, Elon Musk (spot what’s missing from his new xAI team) and Mark Zuckerberg (‘Meta to release commercial AI model in effort to catch rivals’, reports the FT) have been fighting over AI, too.

DSIT up and take notice

So. Farewell then, the AI Council. The Department for Science, Innovation and Technology is apparently ‘establishing a wider group of expert advisers to input on a range of priority issues across the Department, including artificial intelligence’, complementing the Foundation Model Taskforce.

It’s been quite the fortnight for profiles of DSIT’s leadership. Politico’s profile of AI minister Viscount Camrose (the minister for AI, rather than a minister who is AI) included the revelation that the Viscount had never had a Viscount biscuit - now rectified. (There’s a cookie joke in there somewhere.) DSIT perm sec, Sarah Munby, the most senior civil servant in the department, gave an interview to Civil Service World (she wants DSIT to be government’s most innovative department), while chief executive of Ofcom had lunch with CSW.

Aside from that… secretary of state Chloe Smith addressed Chatham House before a panel discussion on global AI governance with Baroness Martha Lane Fox, Microsoft’s Brad Smith and the Tony Blair Institute’s Tony Blair (who also cropped up on Sky News yesterday talking about AI)… Smith also spoke at the Creating a Scientific Superpower Conference … DSIT published details of the UK-Singapore data/tech agreement and its latest monthly cybersecurity newsletter and updated its guidance on ‘secure connected places’ or smart cities… and its start-up board met for the first time (do we get a civil society or academic version?).

And it’s DSIT question time in the Commons next Wednesday.

Parly-vous data?

‘The Government has “no credible strategy” to tackle digital exclusion’ is the stark headline from the Lords Communications and Digital Committee. The Committee says a new strategy and digital inclusion unit should focus on urgent action to help with the cost of living crisis, investment in basic skills, boosting digital inclusion hubs - and ‘future-proofing public services’ to ensure ‘the digitally excluded do not face further marginalisation due to poor representation in the datasets used to inform algorithmic decision making’.

As expected, the Committee has also just launched an inquiry into large language models - you have until 5 September to submit written evidence. If you’re in the consultation-answering groove, you might also want to get on board with the Transport Select Committee’s new inquiry into the ‘future use of transport data’ (you have until 25 August if you’d like to submit evidence) and the Public Administration and Constitutional Affairs Committee’s look at ‘how can Government harness new data to improve policies?’ (you have until 18 August for that one). That comes as the UK Statistics Authority undergoes a review, and as the Office for Statistics Regulation publishes both its annual review of the state of the statistics system (performing strongly, demands for data are becoming more varied, but producers are facing increasing financial and resourcing challenges) and a report on data sharing and linkage in government.

If you care about accurate data then there’s also the Procedures Committee’s report on correcting the parliamentary record, while the European Scrutiny Committee (Commons) quizzed witnesses on regulating after Brexit and the Lords Communications and Digital Committee (Lords) quizzed Ofcom.

And if all that weren’t enough… ‘Is it not extraordinary that we have not previously had a general debate on what is the issue of our age?’, asked Matt Warman MP, kicking off a debate on AI in the Commons.

Labour movement

There are a few mentions of AI and digital skills in Keir Starmer’s speech on the fifth of Labour’s missions, on opportunity, education and childcare. And in a week of leaks in Westminster … you may remember a previous leak of a document full of policy ideas for Labour’s National Policy Forum. There’s now been a leak of amendments to the previously leaked document. Data and tech mentions include an NHS app, online safety and extending Freedom of Information to outsourced public services. Politico have also identified a big tech tax as one of Labour’s recent u-turns, while ‘security’ might be the word to use if you want to influence Labour policy at the moment.

Shadow DCMS secretary, Lucy Powell, gave a speech this week including the choices to be made about who will benefit from AI: “Will it be those who already hold wealth and power, leaving communities behind, as happened with deindustrialisation? Or will technology be harnessed for the public good, building a society where opportunities are widened and the benefits are fairly shared?” (There are similar themes in her contribution to a New Statesman piece on the good digital society.)

The same day, shadow work and pensions secretary, Jonathan Ashworth, talked up the potential benefits of AI, including helping people to find jobs. (He’s also recently spoken about partnering with India on data, tech and AI, and at a Labour Women in Tech event - they’ve been thinking about how more women in STEM would help with Labour’s five missions.) And shadow health secretary Wes Streeting has written a bit about data and IT in a piece for the NHS - helpfully summarised by Tom Forth if you can’t climb the paywall.

There are still rumours of a possible Labour reshuffle, which could include reshuffling to match the reconfigured government departments (including DSIT). Politico has profiled the two most likely contenders for the shadow DSIT brief: current shadow DCMS secretary Lucy Powell, and Darren Jones, chair of Labour Digital and chair of the Business and Trade select committee. He was also recently profiled in The Observer: ‘We have to flip the AI debate towards hope’: Labour’s techno-optimist, Darren Jones’.

Less optimistic was data rights expert Jon Baines, writing about a ‘data grabbing’ Labour website ‘calculating’ what number baby born in the NHS you were.

In brief

Well, it’s briefer than last time:

The UK Government is hiring a chief data officer. Again.
ARIA, the Advanced Research and Invention Agency, is also hiring
The Department for Levelling Up, Housing and Communities has published its plans for Oflog - the Office for Local Government - which is all about data, analytics and the performance of local government
The government says it will consult on a new open data scheme around fuel prices, following a report from the Competition and Markets Authority
The ONS is consulting on the future of population and migration statistics in England and Wales
Data and data-sharing is a key theme of a new NAO report on lessons learned from cross-government working
NHS England’s Transformation Directorate - the artist formerly known as NHSX and NHS Digital - has published an update on how implementation of Data Saves Lives, the health and care data strategy, is going. Jess Morley reckons there have been 21 policy documents with implications for NHS data over the last 18 months, which she’s summarised
There’s a new job as Head of Data Strategy in the new Data Strategy, Platforms & Governance (DSPG) Team in the new Department for Energy Security & Net Zero
Sir Patrick Vallance may have stepped down as chief scientific adviser, but he remains on the board of ARIA, the UK’s new-ish innovation agency
Last time out we noted some ICO work on neurotech - information coming from the brain and nervous system. One (now former) member of the regulator’s technology advisory panel was less impressed
And elsewhere… the Council of Europe has adopted some Model Contractual Clauses to help cross-border data flows, while the European Council and European Parliament have reached an agreement on ‘fair access to and use of data’ (the Data Act).

What we’ve been up to

Adam wrote up our excellent parliamentary event with the TUC on the worker experience of the AI revolution. One of the workers who spoke - actress and voiceover artist Laurence Bouvard, also made it into an FT article on actors losing their voice to AI.
I went along to Open Data Camp in Wolverhampton, and helped run a session on what attendees would like to see on open data in manifestos before the next election. We’re doing some other thinking about what a progressive vision for data, digital and tech policy might look like, too.

What everyone else has been up to

The Ada Lovelace Institute have published a UK-focused policy briefing on what the use of Covid-19 technologies could teach us about future public health crises, and a paper on the role of public participation approaches in commercial AI labs.
The Open Data Institute were up in front of the Treasury select committee about data in policymaking around intergenerational inequality and social mobility
‘2024 will be a litmus test for AI’s effect on elections – and voters’ faith in them’, writes Demos’ Ellen Judson in The Guardian
Jean Innes is the new CEO of the Alan Turing Institute. And the Turing’s Centre for Emerging Technology and Security (CETaS) has just announced it will be providing the secretariat for the All-Party Parliamentary Group (APPG) on Technology and National Security, chaired by Darren Jones
The Institute for Government hosted its 44th Data Bites event, with speakers covering topics including the ONS consultation on the future of population statistics mentioned above
The Child Poverty Action Group published ‘Reap what you code’, an analysis of how Universal Credit ‘adheres to the rule of law principles of transparency, procedural fairness and lawfulness’
Nesta held an event on opening up tech monopolies
The Democracy Network wondered, what’s AI got to do with democracy?
There’s a new Community Tech community of practice
Oxford’s Luciano Floridi is heading over the Atlantic to become Founding Director of Yale’s new Digital Ethics Center.

Events

As mentioned above, Ada have an event with AWO on their report about regulating AI in the UK on 18 July 2023
On the same day, Defend Digital Me are in parliament asking, ‘AI and Education. Existential threat or an everyday toolkit?’ They even have a generative AI poem about it.

Good reads

I’ve been wondering for a while whether senior government figures should be putting important announcements behind press paywalls. The FOI Daily has more on a case involving that very discussion - and the ICO (the complainant has also tweeted about it).
If you want an insight into what can go wrong when we put too much belief in data and digital systems, then Computer Weekly and journalist Nick Wallis are covering the public inquiry into the Post Office’s Horizon IT system and how hundreds of postmasters were prosecuted for accounting shortfalls… that didn’t exist. And on a similar theme…
The Guardian has extensive coverage of Australia’s Royal Commission on the Robodebt scandal, which reported last week, issuing a damning verdict on the Australian government’s use of an automated debt detection and recovery system. Australia’s 7am has a podcast series on the scandal.