Hello, and welcome to our eleventh Data Policy Digest, bringing you all the latest data and AI policy developments.

And Happy New Year! (If it’s not too late to wish you that.) After the data and AI policy carnival of chaos that was 2023, we have a much quieter year to look forward to.

Kidding, obviously. Just the odd controversial bit of data-related legislation, a government response to the AI White Paper, more AI Summits, continued AI development at breakneck speed, the unfolding of the Post Office Horizon scandal, the small matter of a general election being called at some point and goodness knows what else to keep us busy. The Digest will be with you all the way.

As ever, 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 gavin@connectedbydata.org. We’re on Twitter @ConnectedByData and @DataReform. You can also catch up on Digest #1, Digest #2, Digest #3, Digest #4, Digest #5, Digest #6, Digest #7, Digest #8, Digest #9 and Digest #10.

To receive the next edition of the Data Policy Digest direct to your inbox sign up here.

Data policy developments

Deeply DPDIB

Several Lords were a-leaping with anger at the Data Protection and Digital Information Bill when they debated it at Second Reading on 19 December, the last sitting day before Christmas. We’ll have a proper summary soon (in the meantime, you can catch up on my live-tweeting, Big Brother Watch’s summary thread or read coverage ranging from The Big Issue to the Church Times), but suffice to say their lordships were unhappy with many of the same things raised by civil society and in the Commons, including: late amendments including powers for DWP to access benefit claimants’ bank account data, changes to data rules around elections, relaxation of data protection requirements and weakening of rights, threats to data adequacy with the EU, changes to definitions of ‘personal data’ and ‘scientific research’, sweeping Henry VIII clauses investing powers in the secretary of state, and much more.

Labour looked forward to ‘a long and fruitful exchange with the Government over the coming months. This will be a Bill that challenges the Government’, while Lib Dem Lord Clement-Jones warned:

‘We want to amend a great many of the provisions of the Bill and we want to scrutinise many other aspects of it where the amendments came through at a very late stage. I am afraid the Government should expect this Bill to have a pretty rough passage’.

The next port of call in that passage, committee stage, is yet to be scheduled (we hear rumours of mid-February, once the Digital Markets Bill has made it through committee). But it will be a ‘Grand Committee’ - any member of the Lords is welcome to go along, but any changes to the Bill have to be agreed unanimously, meaning most of the amendment action is likely to come through votes at Report Stage. Some amendments have gone down already (and there are some updates to the supporting documentation too).

In other data protection and Bill-related developments:

• Some peers noted concern from the Information Commissioner about ICO was wary of the new DWP powers - he wrote a letter to the Times (h/t Owen Boswarva), who also published a story on it (Computer Weekly have a non-paywalled report)… while Jen spotted the impact assessment on those powers (and some media outlets picked up on the affected banks, the largest 15 in the UK)… and Public Technology reports that ‘DWP signs deal for AI tech to analyse citizens’ social posts in anti-fraud drive’

• Smart data is part of the Bill - the government has just launched a consultation on ‘empowering drivers and boosting competition in the road fuel retail market’ which mentions it, and there was some briefing about smart data in the energy sector to the Mail over Christmas (‘“One-click” energy switch will let households change energy supplier in just seven days under plans considered by Ministers’)

• On the Bill… Stephen Cragg KC summarised his legal opinion for Defend Digital Me (full thing)… there’s a statement from the new Biometrics and Surveillance Camera Commissioner, who will be abolished by the Bill… “Over 35 GLD lawyers from six different departments are leading on measures in the Data Protection and Digital Information Bill”… we held a Connected Conversation on DPDIB and impact assessments … and Big Brother Watch and 38 Degrees have a new petition, Stop the Government from spying on all of our bank accounts …

• And more broadly… the ICO have launched a series of consultations with one on generative AI and data protection … Google are rolling out cookie-blocking within Chrome … and Gloria González Fuster has published an online history of data protection.

Bills, bills, bills

Online Safety Act Many provisions came into force on 10 January, a day after a delegated legislation committee debated the draft Online Safety (List of Overseas Regulators) Regulations 2024 … Ofcom poaches Big Tech staff in push to enforce new internet curbs … there’s a call for evidence as part of the government’s pornography review … and there’s a renewed push around online safety in the US, taking inspiration from the UK approach…

Digital Markets, Competition and Consumers Bill Committee stage kicks off in the Lords on 22 January … the government updated some of the supporting documentation for the Bill… the CMA has published an overview of its provisional approach to implement the new Digital Markets competition regime… there’s more discontent from the right of wonk world about the CMA’s powers… although Microsoft boss, Brad Smith, has toned down his previous criticism of the UK regulatory regime … and the Commons Library has taken a look at the digital services tax …

Other The Automated Vehicles Bill is in Lords Committee stage, with Report scheduled for 6 February… the Investigatory Powers (Amendment) Bill hits Lords Report on 23 January… the Media Bill awaits a date for Report stage in the Commons.

AI got ‘rithm

Some retrospectives: Rest of World have looked at how AI shook the world in 2023, while Marci Harris looks at how the US Congress used generative AI in 2023. Looking ahead, Politico’s Digital Bridge has a preview of 2024, Ethan Mollick (the guy who generated that viral fake of himself) shares some hints about the next year of AI, Stanford HAI asked its experts what to expect on AI and The New PPT its experts about international tech cooperation.

Speaking of things we might expect in 2024… has-the-government-published- its-AIWhitePaper-responseyet.com

There had been rumours the government would publish the response to its AI White Paper consultation this week – perhaps because of Michelle Donelan’s speech (see below). It hasn’t. But the FT have been digging and say there will be ‘tests’ on whether the government should publish new AI laws. “The economic stakes are so high in AI, and, without strong regulatory incentives, you can’t expect companies to stick to voluntary commitments once their market incentives move in a different direction,” Ada’s Michael Birtwistle told the FT.

The government is also launching an ‘AI Opportunity Forum’, co-chaired by Franck Petitgas (the Prime Minister’s Adviser on Business and Investment) and Michelle Donelan to ‘explore in more detail the adoption of AI technologies across the private sector’.

The curtain still hasn’t come down on the EU AI Act drama with debates over wording. Meanwhile, one policymaking participant reminds us of the human cost of such situations… ’EU competition chief defends Artificial Intelligence Act after Macron’s attack’ … ‘on AI, von der Leyen sees glass half full in Davos’ … and a leaked manifesto suggests the EU centre-right is pushing back on red tape around AI.

In other international news… one of the biggest stories over Christmas was around Laion, ‘Largest Dataset Powering AI Images Removed After Discovery of Child Sexual Abuse Material’ (read the original paper)… the FT reports on ‘secret diplomacy on AI safety’ between US companies and Chinese experts… in his new year remarks, the Pope said ‘Artificial intelligence ought to serve our best human potential and our highest aspirations, not compete with them’… the UN’s advisory body on AI has published an interim report, Governing AI for Humanity … the US Government Accountability Office finds US agencies are lacking ‘comprehensive and accurate’ AI inventories … and The Economist takes a look at AI Nationalism …

The UK Supreme Court has ruled that AI cannot be an inventor under UK patent law… the government’s response to the DCMS select committee report on ‘Connected tech’ says it won’t be considering a broad text and data mining exception for AI training… as as thinktank wondered whether the UK could become competitive on text and data mining for AI… the boss of Getty Images told the PM to ‘Back UK creative sector or gamble on AI’ and ‘content creators fight back against AI’ … the New York Times is suing OpenAI and Microsoft over their use of copyrighted work… on the contrary, Microsoft’s CEO says journalists should welcome AI (something something Mandy Rice-Davies)… Press Gazette have looked at ‘tips and tactics’ for generative AI in the newsroom … and BBC Sounds has a short history of copyright …

The IMF says AI could hit 40% of jobs and worsen inequality… the BBC reports ‘Gaming voice actors blindsided by “garbage” union AI deal’ … and the Guardian looks at ‘the pushback against AI surveillance at work’ in the US…

ChatGPT hype obscures the harms AI is already causing … too few of us are talking about certain algorithms… Audrey Tang wonders if superalignment can help AI safety … the FT argues there’smore to AI regulation than safety … a thinktank argues there’s an ‘urgent need for terrorism AI laws’ … the World Economic Forum identifies AI disinformation and misinformation in its Global Risks report…

‘We’re Not Living a “Predicted” Life’ – students give their perspectives on Wisconsin’s ‘dropout algorithm’… young people are turning to AI therapist bots …

The Parliamentary Office for Science and Technology has a new ‘POST note’ on the policy implications of AI … Jerry Fishenden takes a historical view of AI, data, and public services … the LSE says that ‘in shaping AI policy, stories about social impacts are just as important as expert information’ … Dan McQuillan considers ‘AI as Algorithmic Thatcherism’ …

Microsoft got a lot of coverage for introducing a ‘Copilot’ keyboard key … Meta’s chief AI scientist reinvented the library … the Guardian explores a ‘pacy, visually stunning film about the dangers of AI – made by AI’ … the US National Institute of Standards and Technology published ‘Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations’ …

And Happy New Year! ‘Could an AI ‘death calculator’ actually be a good thing?’

DSIT up and take notice

As well as a nice Christmas bit from the DSIT social media team, the end of 2023 brought various round-ups of the year in government data and digital, including from Public Technology’s top stories, parts one and two … Computer Weekly’s top 10 government IT stories (and other compilations, including the top AI regulation stories)… Global Government Forum’s top news stories, webinars and features …

There were look-aheads from key government data and digital figures, with DSIT’s perm sec wanting the department to allow ‘people to think big, take risk, and move fast’ … the head of GDS’s top priority being around One Login … the national statistician looking to 2031, and whether there’ll be a traditional census… a range of leaders talking to Government Transformation Magazine about their priorities… the head of the National Audit Office ‘put(ting) digital and AI in the frame for ongoing modernisation’ (he said similar to parliament this week)… and some general trend-spotting from Public Technology and UKAuthority …

And there were some new year’s honours for tech people in government including the CDIO at DSIT and a Number 10 data scientist. OSB campaigner Ian Russell was among the others honoured.

Michelle Donelan’s speech this week didn’t launch the government’s AI White Paper response. But it did set out some of DSIT’s priorities for ‘scale ups’, including setting up a forum and new regulatory support service, backing a new Regulatory Horizons Council report (they’ve published another on moving from regulatory principles to practice too, UKRI have also published something), and ‘making data access a government priority this year… whilst of course continuing to prioritise data privacy and data consent’ given the ‘potential goldmine’ of data central and local government holds. (This was true back in December 2022 when the government promised a review of its open data policy, which you can find here.)

There’s been lots on AI in government (just for a change)… DSIT perm sec, Sarah Munby, told parliament that 6% of the department’s officials are ‘completely dedicated to work on AI’… and the AI Safety Institute is recruiting … Chancellor Jeremy Hunt highlighted the UK’s ‘light touch’ regulatory regime as he ‘championed’ British tech when he spoke on a panel including OpenAI’s Sam Altman at Davos … deputy PM Oliver Dowden spoke about the opportunities for AI in government, while the minister for the Cabinet Office talked to The Times … Alex Burghart, the Cabinet Office minister responsible for AI, spoke at the Centre for Policy Studies about using AI in government, generating headlines on not outsourcing AI projects to tech firms, replacing civil servants in fraud detection jobs, and using AI for institutional memory while proceeding with caution… a Cabinet Office algorithm is already helping to sort and delete files (and details have been published under the Algorithmic Transparency Recording Standard)… GDS have been experimenting with using generative AI to help users of GOV.UK … DSIT’s AI lead in the Commons, Saqib Bhatti, warned the ‘fabric of society’ hangs in the balance of AI regulation (and spoke to Computer Weekly, mainly about chips)…

In other ministerial news, Julia Lopez returns from maternity leave, replacing John Whittingdale (she’ll be responsible for DPDIB when it comes back to the Commons). And in other personnel news… AI Summit sherpa, Jonathan Black, is returning to government as deputy National Security Adviser… while James Snook, former head of data policy at DSIT and DCMS, has joined Flint Global.

In other news from DSIT… the UK and Japan agreed some joint tech priorities … a couple of new Turing fellows have been appointed … we’re invited to help shape the new National Maths Academy … and you can check out who ministers (and their special advisers) were meeting and receiving hospitality from between July and September… the agreements between DSIT and its arm’s-length bodies, and Ofcom, DSIT and the Welsh Government … and letters to senior responsible officers on some of DSIT’s major projects.

From elsewhere in government… the civil service’s new People Plan has bits and pieces on data skills and literacy… government decisions made via WhatsApp must go on the official record … the Department for Business and Trade have launched a new tech programme in partnership with Ukraine … and ‘departments’ funding woes could hamper £229m data-sharing platform’, says the national statistician.

Parly-vous data?

The first Wednesday back in the Commons saw DSIT questions and an urgent question on the Post Office Horizon scandal as the bread in a PMQs sandwich. DCMS questions the following day also had one on AI and the creative industries. Also in the Commons, the education committee quizzed experts on ‘how should children be taught about online harms and digital literacy?’ and the Public Accounts Committee covered some data and digital issues with civil service chief operating officer, Alex Chisholm (who steps down in a few months). There was also a Westminster Hall debate on ‘Online Filter Bubbles: Misinformation and Disinformation’.

In the Lords, the DPDIB debate was preceded by a question on confidential medical records and the NHS app, and the Communications and Digital Committee has published some new evidence for its LLMs inquiry, including from OpenAI (on using copyrighted material) and DSIT (which touches on the controversy about whether the central risk function controversy, a saga regular readers have been gripped by).

Labour movement

Keir Starmer made a new year’s speech which named technology as one of the big challenges of the age… the party has published a single document bringing together its missions, and there are reports manifesto policies need to be finalised in February … one of the mission areas is education: shadow ed sec Bridget Phillipson gave a speech which touched on their children’s number pledge; Jen Persson has written more about data in schools off the back of it, while medConfidential argue that ‘NHS numbers are for life, not just for children’ … the Welsh leadership contest is in full swing, with opening pitches from Jeremy Miles and Vaughan Gething not having a huge amount to say on data, tech or AI…

Politico (paywall) reports shadow DSIT secretary, Peter Kyle’s, new advisers will start soon (and that he’s accused the government of inconsistency in tech policy, and has been meeting techUK members)… Kyle also talked about data centres and planning reform (47 minute mark) - Google yesterday announced it’s investing in a new data centre in Hertfordshire … there ‘s been some briefing about putting tech at the centre of public service reform … and some Labour research on the scale of government legacy IT systems at ‘critical’ risk.

On the Horizon

One tech story has dominated the news over the last few weeks, with the Post Office Horizon scandal (finally) breaking through into wider public consciousness thanks to an ITV drama, Mr Bates vs the Post Office. There’s plenty out there - Computer Weekly (one of the key outlets in reporting the story) have an ‘everything you need to know’ piece; the BBC have made their vital Panorama of a few years ago available; and I can heartily recommend Nick Wallis’ book on the scandal - his website also has links to Private Eye and other reporting.

Some pieces we’ve seen floating around our networks… a set of lessons from IfG which includes a few paragraphs from me on the tech/human angle… mySociety have a summary, including the importance of FoI in bringing facts to light… Peter Wells wrote a few months ago about the computer evidence angle, also the subject of work by City University, James Christie and David Allen Green … the FT and Bloomberg reflect on the role of tech… the Open Government Network’s Kevin Ketih took an open government angle with The Independent… former head of the Government Digital Service, Mike Bracken, and former Blair adviser, Geoff Mulgan (you may have seen his prescient 1999 memo doing the rounds) shared their perspectives… while the Royal Statistical Society reminded us of another tech system-driven scandal, Robodebt in Australia, and ‘DWP boss “really hopes” fraud-busting IT scheme doesn’t echo Horizon scandal’, reports the Disability News Service.

The Inquiry continues.

In brief

• Let’s see if you can distinguish Malevich’s famous Black Square painting from the government’s Federated Data Platform contract with Palantir … Bloomberg reports that the NHS Probes Whether Palantir Campaign Breached Contract Terms (the Good Law Project have published something on that too) … as minister says the FDP ‘will be more secure than anything currently used in NHS’ … and DHSC signs ‘ministerial intelligence’ deal with X to analyse posts by would-be hires (Civil Service World)

• Police to be able to run face recognition searches on 50m driving licence holders, reports The Guardian

• Airbnb, eBay, Vinted and other apps are to share more information with HMRC … while TikTok’s data collection being scrutinised by Australia’s privacy watchdog

• The British Library will have to ‘burn through reserves to recover from cyber attack’, reports the FT… as the BL updates on restoring its services and the BBC takes a look, too

• A new Public Communication Scan of the United Kingdom from the OECD has stuff to say on data, stats and public engagement.

What we’ve been up to

• We’ve been planning an event! On 13 February, we’re holding an online event, ‘Datafied’: How data and AI policy will impact all areas of progressive action, and what civil society can do about it? This is a free virtual workshop for civil society and campaigners to get to grips with how data and AI affects your areas, whether that’s education, health, benefits, housing, social care, work, policing, democratic engagement or a multitude of other areas. The session will give an accessible introduction to those with a general level of knowledge of the issues, focusing on how current data driven and AI systems affect people and relevant policy areas, provide an opportunity to discuss campaigning approaches on those issues, and take a practical look at the new data legislation going through Parliament and how data and AI policy issues are likely to feature in the coming General Election. Come, and please do spread the word to your networks!

• Maria blogged about G20 Brasil: an opportunity to shape how we shape technology (and summarised a couple of webinars on Data Governance and Artificial Intelligence in Mercosur)

• Jeni, Emily, Tim, Adam and Maria all reflected on 2023

• We’ve published a summary of our People’s Panel on AI from the AI Fringe, including video - and you can also watch our session at techUK’s Digital Ethics Summit

• We’ve summarised our recent Connected Conversation on DPDIB and impact assessments

• Judith Townend, one of our fellows, has a new book out on open justice and transparency in the criminal courts… and Natalie Byrom, another of our fellows, has written about how ‘Unregulated AI legal advice puts the public at risk’ for the Law Society Gazette (picking up on guidance published just before Christmas) and wrote for the FT on how ‘Victims are getting lost in the justice system — IDs could help’.

What everyone else has been up to

• Artists fighting back against AI with data poisoning and the need to fight for cognitive liberty and neuroprivacy are among Nesta’s future signals to watch in 2024, while ‘digital’ is one of three key areas for national security and industrial strategy, according to a new Labour Together report

• Ada’s Matt Davies featured in a New Statesman piece on 2024’s key questions for Labour, as Ada’s director and interim director looked back at 2023

• Lots from the Open Data Institute over the break, starting with a new data-centric AI project - ‘What do we mean by “without data, there is no AI”?’ … their chair, Sir Nigel Shadbolt, wrote for the Evening Standard on ‘why we need to fear the risk of AI model collapse’ and for City AM on why ‘preventing the AI dystopia starts with good data’ … and away from AI, a new report, ‘From concerns to consent: addressing the issue of trust and other challenges impairing secondary use of health data in Poland and beyond’

• Big Brother Watch’s Silkie Carlo wrote ‘Facial recognition at airports might sound convenient for travellers – but it is putting us at risk’ for the Sunday Telegraph

• Chatham House published ‘Towards a global approach to digital platform regulation’

• Wired: ‘A Damning Report Claims the Union That Defeated Uber Had a Culture of Abuse and Toxicity’

• mySociety have a report celebrating the one millionth FoI request made via their services, Freedom of Information to support social change

• The Alan Turing Institute have a new study, Generative AI is already widespread in the public sector

• Public Law Project have a new report, ‘“Constantly on Edge”: The expansion of GPS tagging and the rollout of non-fitted devices - an annual review of GPS tagging in the immigration system, 2023’

• There’s a new initiative to ‘help advance the UK’s science & tech policy landscape’ called UK Day One

New year, new job? Some we’ve spotted are Tech Policy Adviser at Labour Together … Senior Advisor - Data & AI at TBI … a few on AI and AI compliance at the ICO … and various roles at Public Law Project … while Elizabeth Seger will join Demos as director of their Centre for the Analysis of Social Media (CASM) in February

Events

• One from us to kick off - our online workshop, ‘Datafied’: How data and AI policy will impact all areas of progressive action, and what civil society can do about it? is happening on 13 February

• Machina Judicata: Automation, AI and Justice - views from the court, organised by Brick Court Chambers, leading AI law firm AWO, and UCL Laws, is happening today

• Demos have a launch event for a report on ‘Generating Democracy - AI and the coming revolution in communications’ on 23 January

• Open Data Manchester have a series of workshops on putting people first in AI, kicking off next week

• On 1 February, Public Law Project have a workshop, Rise of the Robots: The harm behind the hype

• State of Open Con 24 takes place in early February

• The IfG’s 48th Data Bites on 26 February is a smart data special

• mySociety’s TICTeC (The Impacts of Civic Technology Conference) returns in June, and is happening in London

• On 11 March, techUK are holding their ‘Tech Policy Leadership Conference 2024 - How can the next Government use technology to build a better Britain?’ with speakers including Michelle Donelan and Peter Kyle

Good reads

• Forget technology — politicians pose the gravest misinformation threat (FT)… Fears UK not ready for deepfake general election (BBC)… and check out the data-driven campaigning bit in Digital Bridge (Politico)

• Why Linda Yaccarino took on the wildest job in Silicon Valley (FT)

• Democratising data - how Fifa primed World Cup underdogs to shock (BBC Sport)

• Who Is Killing All These Stories About a Controversial Tech Mogul? (Daily Beast)

• Big Tech has already made enough money in 2024 to pay all its 2023 fines (Proton)

• Kafka in the Age of AI and the Futility of Privacy as Control (Daniel Solove and Woodrow Hartzog)

• YOU CAN CALL ME AI (Bill Thompson)

• The Disturbing Impact of the Cyberattack at the British Library (New Yorker)

• Is the West talking itself into decline? (FT) - note the various responses and quote tweets

• The content moderators fighting for their rights (FT)

• Do No Harm Guide: Collecting, Analyzing, and Reporting Gender and Sexual Orientation Data (Urban Institute)

And finally:

• Our data problems are getting harder to ignore (The Straits Times, via Quantum of Sollazzo)

• The BBC may have found a rare exception to Betteridge’s Law over at the BBC: ‘CES 2024: AI pillows and toothbrushes - is it all getting a bit silly?’