My week
Monday we had some of our normal weekly meetings, in particular discussing our approach to the Data Bill and the write up of the design lab we ran a few weeks ago looking at worker and public participation in public procurement of data and AI systems, which Adam is working on. I also had an interesting chat with John Pullinger about the role of the Turing Institute in the public policy space, and a catch up with Jack Hardinges.
Tuesday I worked on a concept note for a design lab we’re planning with the Office for Statistics Regulation, to put together some advice about how to involve the public in the design of official statistics. I caught up with our fellows, including a discussion about how to engage children about issues to do with data and AI. Then I headed into London to catch up with Tom Smith at MHCLG, and go to the ICO 40th birthday event.
Wednesday I spent the day at the Microsoft offices at a Charities AI War Room organised by CAST and Zoe Amar Digital, where we discussed how charities should respond to AI.
Thursday I had at home; aside from a couple of meetings I mostly worked on a briefing for the Lords Committee stage of the Data Bill, focused on amendments around recognised legitimate interests.
High points
- Listening to our fellows, at various different stages in their careers, support and advise each other.
- Hearing one of the charities at the War Room on Wednesday, who had only been thinking about how charities adopt AI themselves, reflect that he wanted to think more about how AI was going to affect their beneficiaries, and what that meant for their mission.
- Seeing different civil society groups debate joint tactics in our approaches to amendments to the Data Bill as it goes through different stages in the Lords.
Learning points
- The speeches at the ICO 40 event were incredibly powerful. I’m carrying two things from it. One is the important use of subject access requests to get hold of files kept on people (particularly by social services) both to hold the system to account and to gain their own resolution with their past. The second is the awful impact of Stop and Search on Black boys and men.
- I really felt the seductive power of the “AI is progress” narrative during the charities workshop and how easy it is to get swept up in a fear of being left behind – of being on the wrong side of a growing adoption and productivity divide. I like the spheres of influence diagram that Dan used to describe different levels of impact, but I do think we need to critically appraise the framing of gaining “efficiencies” and consider what might also be being lost.
- Apparently, the last time a statutory instrument failed to pass the affirmative procedure was in 1978. The last time one failed to pass the negative procedure was in 1979 when the Government failed to shout “No” at the right time.