Hello! Or rather, tere!
Our sixth Data Policy Digest, bringing you all the latest data and AI policy developments, comes to you from this year’s Open Government Summit in Tallinn, where Tim, Helena and I are flying the flag for Connected by Data. We ran a workshop yesterday; give us a shout if you happen to be here in Estonia.
If not, there’s plenty to catch up on back in the UK, as parliament returns after its summer break.
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 email@example.com. We’re on Twitter @ConnectedByData and @DataReform. You can also catch up on Digest #1, Digest #2, Digest #3, Digest #4 and Digest #5.
Data policy developments
We’re waiting on a date for Commons report stage - that’s when the Bill, having been examined by a committee, comes back for MPs to consider amendments. Given MPs will only be around for a few weeks before conference recess, and DSIT has to get the Online Safety Bill through before this session of parliament ends (some time before the next State Opening, on 7 November).
We know some of the amendments that might be up for discussion, though - there’s one on exempting the police and Crown Prosecution Service from certain data protection principles (devotees may remember amendment mover, Jane Hunt, raising this during the second reading debate) and another on Smart Data. John Penrose, who put forward that amendment, recently wrote for City AM on the topic.
Elsewhere … our friends at the Open Rights Group did a quick video explainer … APIL, which campaigns on behalf of people injured by negligence, wants an amendment to ban cold calls about making personal injury claims … openDemocracy looks at the possible impact of the Bill on protecting health data … Politico looks to the US, where ‘one of the world’s largest advertising firms is crafting a campaign to thwart a California bill intended to enhance people’s control over the data that companies collect on them’ - and the BBC looks to India, with an article on the country’s new Data Protection Bill.
Bills, bills, bills
Online Safety Bill DSIT has published an updated guide to the Bill.
Forget any of the recent big boxing bouts - the real heavyweight contest has been between Lord Bethell and Signal’s Meredith Whittaker, after the former accused the latter of ‘unpersuasive catastrophising’. Whittaker is one of those who spoke at a civil society event in parliament yesterday.
The Guardian reports that Australia - which had been following a similar approach to the UK - ‘will not force adult websites to bring in age verification due to privacy and security concerns’. BCS says the OSB shouldn’t rely on technology to deliver child protection and Apple has issued a statement to say ‘it could not proceed with development of a CSAM-scanning mechanism, even one built specifically to preserve privacy’.
And in a select committee report on Henry VIII powers (clauses in legislation which allow the government to fill in the detail afterwards), the Lords are Boleyn with rage. Click here if you’d like to Seymour. (Terrible puns, even about Henry VIII’s wives, are Parr for the course in this Digest. Howard you cope without them?)
Digital Markets, Competition and Consumers Bill We looked at the regulatory rigmarole around Microsoft’s attempt to buy games giant Activision Blizzard back in Digest #4. There were many questions about what the UK’s Competition and Markets Authority - which was concerned about the deal lessening competition - would do after the EU said it was satisfied and a judge ruled against the US Federal Trade Commission’s objections. Well… the CMA stood its ground and Microsoft has had to restructure the deal accordingly.
Elsewhere, UK in a Changing Europe carries a piece on what Meta’s decision to keep its Threads social media platform (remember that?) out of the EU market by Anu Bradford, who has a book out on ‘the global battle to regulate technology’ in a few weeks. And 56 organisations have called on the European Commission to use its powers under the Digital Services Act to safeguard Europe’s 2024 elections.
AI got ‘rithm
We have a date! For the Summit!
And we have some details! DSIT says that, following an ‘iterative and consultative process’, building on ‘initial stakeholder consultation and evidence-gathering’, there are five objectives: a shared understanding of the risks posed by frontier AI, a ‘forward process’ for international collaboration on frontier AI safety, appropriate measures for individual organisations to take on frontier AI safety, areas for potential collaboration on AI safety research, and showcasing how the safe development of AI will enable AI to be used for good globally. (Note that frontier AI is a term coined by OpenAI.) DSIT also says that attendees will ‘bring together key countries, technology organisations, academia and civil society’.
Despite that, we’re still concerned about civil society and public voices being excluded - there was a glimmer of hope as DSIT mentioned civil society in a tweet (and first in the list, no less), and then trumpeted a call with a small group of civil society figures (also highlighted in yesterday’s announcement. It remains to be seen if any invitations will only be a fig leaf, or we’ll actually be in the room where it happens (or at least, the Zoom where it happens). Turns out start-ups are also concerned their voice isn’t being heard, as one venture capital firm urges the UK government to stick with its white paper approach to AI regulation ahead of the Summit.
Voices from the Global South should also be included, but are currently missing from the AI regulation debate. On the international front, speculation continues as to whether China will be invited - Politico reports there will be some involvement, a sense strengthened by the foreign secretary’s recent visit to Beijing. Meanwhile, BRICS - which includes China - have agreed to establish an AI study group for information exchange & cooperation. In the US, Senator Chuck Schumer is hosting his own summit dominated by US tech figures.
Women’s Hour and Brookings are among those thinking about AI’s impact on jobs, while The Observer has more on our friends at the TUC and their taskforce examining the threat to workers’ rights.
A couple more AI white paper responses for you: the Law Society say there is ‘an urgent need for explicit regulations delineating liability across the AI lifecycle, and for guidelines which clearly describe under which circumstances an entity may be held liable for the outcomes of an AI system’, while I wrote up a workshop with Open Data Institute members.
In UK government news: the National Audit Office has a new project on the use of AI in government … the ICO and Digital Regulation Cooperation Forum discussed AI at an IAPP event … and the ONS blogged about how LLMs might help with finding stats.
Elsewhere: Spain has created an AI regulator… CIGI have looked at the Canadian approach … a lot of people have been sharing a new paper on what ‘open source’ means in relation to AI, a subject that Wired also examined … Timnit Gebru spoke to Ethan Zuckerman… Deepmind’s Mustafa Suleyman spoke to the Leading podcast and The Guardian about his new book … the FT explores the sceptical case on generative AI … Kriti Sharma, founder of AI for good, is worried about ‘AI fear-mongering’ … The Economist is less worried about the impact of AI-generated disinformation on elections than you might think… and Google is testing a watermark to identify AI images, as it celebrates its 25th birthday.
DSIT up and take notice
Various bits and pieces from DSIT and its associated bodies over the last few weeks, including: backing for AI projects to help grow the clean energy economy and a new Research Ventures Catalyst (now open for applications)… UKRI is cashing in on chips … there’s a vocal warm up ahead of ARIA announcing its programme directors … the Geospatial Commision has delivered a new commissioner (the chief data officer at Ocado) and a new report on data and electric vehicle charging points … and mein GOTT! The Government Office for Technology Transfer has published new guidance to develop a Knowledge Asset Management Strategy.
Over at Cabinet Office, the Data Standards Authority has published a catalogue of standards, and GDS have blogged about the new in-person identity check for GOV.UK One Login and building a green digital service.
As well as the DCMS committee’s report on AI and creative technology (see the AI got ‘rithm section), we have an interim report from the Science, Innovation and Technology Committee’s inquiry into AI governance. It identifies 12 challenges of AI governance that policymakers need to address: Bias, Privacy, Misrepresentation, Access to Data, Access to Compute, Black Box, Open-Source, Intellectual Property and Copyright, Liability, Employment, International Coordination, and Existential. (Interesting that ‘Environmental’ isn’t one of them - we noted a fair bit of coverage around that in our last Digest.) Labour’s then-shadow digital secretary, Lucy Powell (see below) liked it.
We have a good few weeks to wait before the next DSIT question time in the Commons - nothing in the latest calendar which runs to 26 October - but you can hear National Statistician, Sir Ian Diamond, give evidence to PACAC’s inquiry on the UK’s evidence base today. Meanwhile, the Commons Library has a briefing on AI and employment law.
We might also expect a deluge of answers to written parliamentary questions as the Houses return. Hopefully some will be more informative than this one on the use of automated decision making in the Department for Business and Trade.
To quote the great Bart Simpson: ‘I don’t think any of us expected him to say that’.
After months of speculation, the Labour reshuffle finally happened. Lucy Powell – previously shadow secretary of state for digital, culture, media and sport – has moved to become leader of the House of Commons. Darren Jones – previously chair of the Business and Trade select committee, tipped by many to take shadow DSIT secretary, and author of a recent piece on what he wanted to see from the AI Summit – instead became shadow chief secretary to the Treasury. (Not actual chief secretary to the Treasury, as Jones tweeted – though having someone in that role who understands data and digital could have a big impact in what gets funded and how Treasury funding processes operate.)
The new shadow secretary of state for science, innovation and technology is… Peter Kyle. In his tweet about it he says he wants to foster innovation, strengthen our tech sector and ensure science fuels growth and prosperity (though click here for a surprise). Kyle has been MP for Hove since 2015 and was shadow secretary of state for Northern Ireland – he previously held junior shadow roles on schools, and on victims and youth justice. He was previously a member of the business select committee, and the Labour Party’s backbench committee on business. Before politics, he worked in the charity sector, and he has a PhD in community economic development.
At time of writing, it doesn’t look like junior frontbench positions have been confirmed: both the Labour Party website and LabourList (which has lots of reshuffle coverage) list only Kyle under DSIT, suggesting there may be changes to come (which could simply be rebadging jobs already held by MPs such as Chi Onwurah, Steph Peacock and Alex Davies-Jones, or bringing new people in). For those interested in how a Labour government would approach data and digital, there have also been a few appointments to the shadow Cabinet Office team – chancellor of the Duchy of Lancaster Pat McFadden, paymaster general Jonathan Ashworth and minister without portfolio Nick Thomas-Symonds – though it’s not yet clear who will lead on what.
Elsewhere … we are no longer waiting for Sue Gray to start as Keir Starmer’s chief of staff (she did so yesterday). The Times reported that she will have a hand in thinking about how to rewire Whitehall (Civil Service World have a non-paywalled write-up), and the New Statesman suggests we might soon get some more detail on what Labour’s missions ‘mean for the machinery of government and the digital state’.
On the face of it: a couple of controversial facial recognition stories at the Home Office, which has launched ‘a new Market Exploration which seeks technological solutions for the use of Facial Recognition (FR) technologies within policing and other security stakeholders’ (written up by the FT) and is accused of lobbying on behalf of a facial recognition company (first noted by data protection expert Tim Turner)
Health check: the National Data Guardian has written about the NHS’s procurement of a Federated Data Platform and the need to engage the public (Understanding Patient Data and our own Jeni shared their thoughts, too, while I wrote up a roundtable on what the NHS should learn from GPDPR earlier this year), with further reports from openDemocracy
Data migration: the Guardian reports concerns that migrants will get Home Office reference numbers on their NHS England records, and looks at the Home Office’s historical interest in migrants’ health records
From cops on dots to leaks on beaks: more news of police data breaches (dear friends, once more). The Northern Ireland Affairs committee is taking evidence on the PSNI breach on Thursday. And though less digital in nature confidential files found at a derelict police station.
Target practice: The Observer report that the Government is ‘targeting UK minorities with social media ads despite Facebook ban’
And… the ICO have looked at customer data being used for unwanted romantic contact and made a statement on data scraping and data protection … the Office for Statistics Regulation is undertaking a review of the Code of Practice for Statistics ‘to ensure it continues to effectively serve the production and regulation of government statistics’… and DLUHC have published a report on barriers to sharing data with local government.
What we’ve been up to
We published our first annual report!
What everyone else has been up to
The Ada Lovelace Institute and Nuffield Council on Bioethics published DNA.I., a report of early findings and emerging questions on the use of AI in genomics
The Centre for Emerging Technology and Security (CETaS) at The Alan Turing Institute and Centre for Long-Term Resilience have published a new report, Strengthening Resilience to AI Risk: A guide for UK policymakers
New projects galore: from the Bennett Institute (Cambridge) on AI and geopolitics; from the Startup Coalition (the artist formerly known as Coadec), Onward and the Tony Blair Institute on AI; and from the Royal Statistical Society on ‘statistics under pressure’ (to support statisticians working under pressure to feed data into decision-making)
Daron Acemoglu has written Reframing AI: from Machine Intelligence to Machine Usefulness for IFOW
Robin Carpenter has written about medical AI, specifically: How do I get my medical AI into an NHS hospital?
Party conference season will soon be upon us! Connected by Data will be at the Green Party conference (more details to follow) and in Liverpool for the Labour Party Conference. At Labour, we’re holding a discussion with our friends at the Fabian Society on what any Labour government should (or shouldn’t) do on data and AI policy in their first 100 days. We have a brilliant panel - as well as us and the Fabians, we have Mary Towers from the TUC, Mathew Lawrence from Common Wealth, Shameem Ahmad from Public Law Project and Tom Adeyoola from Labour’s start-up review - and we’ll be encouraging lots of contributions from the audience. Join us at fourth floor gallery, Tate Liverpool, 1730-1900 on Monday 9 October (registration link coming soon).
Speaking of IfG… their latest Data Bites event is happening at 6pm on Wednesday 13 September
That same evening, Demos are in conversation with shadow tech minister (at time of writing), Alex Davies-Jones
The Political Studies Association’s Technology, Internet, & Policy Specialist Group has a roundtable on AI regulation and responsible innovation
Innovate UK-funded Bridge AI are holding several residential innovation labs ‘to co-create game-changing AI-powered solutions and increase business productivity in Agri-food, Construction, Creative Industries, and Transport sectors’
And… this year’s Royal Institution Christmas Lectures will be given by Prof Mike Wooldridge on AI.
Elon Musk’s Shadow Rule - The New Yorker
‘Business deserves deeper insight into the security threats the UK faces’, where geopolitics includes AI and digital technology, cyber warfare, trade and green technology, writes a former foreign office perm sec in the FT
As Militaries Adopt AI, Hype Becomes a Weapon - JSTOR Daily
How robotaxis are dividing San Francisco - BBC News
The rise of the tech ethics congregation - MIT Technology Review
- The tech angle on the Lionesses’ World Cup run.