Lots of update in these, rather sporadic week month notes.
East Coast and West Coast
In the middle of September I had a week in the USA, anchored around a draft paper I was presenting at the World Bank Global Forum on Coalitions for Reform (still in review, more on that soon hopefully), but building in the opportunity to catch-up with friends and colleagues in Ithaca, Washington, New York and the Bay Area.
Thanks to kind hosting by Joe Foti at the Open Government Partnership I was able to run a hybrid workshop on draft principles for public participation in AI procurement (see below) at the Open Gov Hub, and the timing of the trip worked out just right to be able to join the Open Environmental Data Project for a convening on climate data infrastructures on Oakland (hosted in a building just next door to the home of one of my favourite infrastructure/architecture podcasts).
I was also able to catch up with some of the team from the Collective Intelligence Project, and with leaders of the upcoming industry-wide deliberative forum on AI Agents, hosted out of the Deliberative Democracy Lab at Stanford: helping shape to some of the ideas in this tech policy press piece I co-authored with Anna Colom, building on the launch of the UN’s Global Dialogue on AI Governance, and reflecting back on the report we launched this time last year on options for deliberative governance of AI. I also caught up with Roya Pakzad of Taaraz, on some of the impactful research she has been leading on technology & human rights.
Education, Education, Education
Back in the UK I’ve been working on the kick off for a small new contract we are working on for the Department for Education to develop lesson and workshop resources to support student dialogue around generative AI in schools.
The goal is that, by the end of October, we’ll have a small set of resources that help teachers, youth leaders or facilitators to run sessions that both introduce how generative AI is being, or could be, used in classroom and school contexts - and that support students to consider different pro- and con- perspectives on this - before putting forward their own views, and messages to the people building and shaping AI.
Last week I was digging into different AI teaching tools, looking at the literature on AI in education, and trying to see what’s possible to fit into a single or double 45 minute lesson plan! Next week I’m hoping to get some early draft resources put together.
I’ll be looking for a few reviewers to take a critical look at the resources in late October, so if that might be of interest, do get in touch.
This work builds on commitment that Department for Education have shown to student and wider public engagement as they develop their AI approach, coming after a recent exercise with legislative theatre they ran with Thinks at the end of August. I had the opportunity to go along to the final presentation of that work, seeing how secondary and college students were reflecting on the potential, and pitfalls, of student data being collectively licensed to support AI model training.
PAIRS prepares for India
It’s also been a busy month for the Participatory AI Research and Practice Symposium as we launched our Call for Abstracts for a hybrid event with online sessions planned for 17th Feb 2026, and in-person sessions planned for 18th Feb 2026 alongside the AI Impact Summit in India next year.
We’ve been having regular programme committee meetings. We’ve also just had exciting news about funding for our work on PAIRS over the next year that means we should be able to start thinking also about various PAIRSx events: thematic spin-off convening looking at how participatory AI practice is developing in various specific sectors or regions (e.g. education, humanitarian, health).
Can Open Government Address Digital Power?
I’m writing this as I head back from the Open Government Partnership Global Summit in Vitoria-Gazteiz. I was there running a fringe event with ParticipationAI, Open Contracting Partnership and Posterity Global to launch the Principles for Public Participation in the Procurement of AI (now available at P4AI.net), and to facilitate a fishbowl conversation on participatory governance of AI.
The launch workshop for the principles was really encouraging: with a packed room and great engagement. We trialled an adapted version of the Good Governance Game and had discussions about potential next steps for the principles.
In the coming weeks I’ll be catching up with the ParticipationAI team to think about how we might fundraise for further work to support a number of local or national governments to trial putting them into practice, and to generate case studies and examples of how participatory procurement can work in practice (get in touch if you have ideas, or are interested!).
As it was, I think, my 8th global OGP Summit (OGP has been the genesis or a key locus in many of the past-projects I’ve worked on, from the Open Data in Developing Countries Research network, to the Open Contracting Data Standard, Open Ownership, Global Data Barometer, Open Data Charter Open Up Guides, and State of Open Data book), I was reflecting a lot on both the ongoing strengths of a community working to improve governance in tough times, but also the challenge of adapting and updating our strategies and understanding to better confront the anti-democratic forces of inequality, and the growing power of tech-oligarchs.
In this context, I have a sense we have much more work to do in order to develop the narratives that can help the democracy and governance field engage more directly with technology-related forms of power. In a breakout session focussed on data centres if felt like we were able to get closer to these conversations: as the material and neo-colonial realities of vast data centre construction projects competing for natural resources of energy and water is more legible to a field that has thought lots about how we govern extractives industries. But, when main stage presentations and side-discussions of civic technology continue to treat AI as an abstracted set of apolitical tools, and few of the OGP Challenge Commitments are grappling with governance of digital technologies, there are gaps to address. This reflects some of the findings the mapping study we did for OGP earlier this year, and that I’ve finally managed to share in our resources section.
All that said, I didn’t get into all that many of the plenary or side sessions as there were many people, personal and professional updates to catch-up with the hallways.
Houses, houses, houses
This month we also wrapped up our small contract for Peabody Housing, looking at resident attitudes towards the new data strategy for a housing association managing over 100,000 properties. As, because of the project timeline, we ran most of our engagement activities over the summer, we spoke with tens, rather than hundreds of residents. But it was striking that even this generated some powerful insights to inform the framing and focus of the data strategy: and it was good to be able to present this back to the team who’ve also been engaging with staff and other stakeholders.
Bold politics
The last highlight to mention from this month was getting to spend a day and a half at the Green Party of England and Wales conference. I was there to facilitate a meeting of the re-energised Science & Technology Policy Working Group that I’m interim co-convenor of with Graham Tavener, though I was also able to stay for the inspiring leaders speeches, some of the voting plenaries, and to join the Red Line for Palestine action on Bournemouth beach.
When we stewarded the through an AI policy for the Greens, we had a relatively small group of interlocutors inside and outside the party. Now it feels like many more people are now interested in discussing and shaping Green science and technology policy: and it’s been interesting to think about how we can develop a process that supports informed deliberation on topics like digital ID, online safety and data governance - in the context of a growing party where we have specialised experts, beginner generalists, and everything in-between - and where so many policy areas are interlinked.
I’m mostly working on this in voluntary time outside Connected by Data, though it’s providing a really useful canvas for thinking about different forms of collective dialogue on data and AI policy, and the value of up-to-date and accessible explainers for some of the technology policy decisions we face.
What I’ve been reading?
- The Global Index of Responsible AI have published a thematic feature on participation, looking at indicators and examples from across the world of where the public are being given the chance to engage with AI policy.
- I’ve been catching up with NESTA’s AI Social Readiness Advisory Label project which sets out a detailed deliberative polling process to evaluate AI programmes, testing it on the UK AI Incubator’s (i.AI) Consult tool for AI-driven analysis of public consultation responses. The pilot report provides detailed feedback for the Consult tool developers based on informed small group discussions. It’s encouraging to see responses in the report itself from the tool developers on how they have taken on board public feedback.
- A few weeks back I finished reading Karen Hao’s Empires of AI: a brilliantly researched and compellingly written account of the development of OpenAI, from early non-profit days, to the hyper-scaling disruptive firm of today. Wholly recommended, and the chapter on data centres has informed many of the discussions I entered into on that topic over the last month.
- Karen Hao’s book was also useful context for reading the Open AI Non-Profit Commission’s Report which, although fatally limited by failing to engage with groups beyond the United States during it’s development (particularly when OpenAI data released last month showing many countries have much heavier use of ChatGPT than the USA), provides a relatively robust account of how Open AI could live it’s purported values. And yet, the $50m grant round announced alongside the report betrays a seeming wilful failure to engage with the report’s headline recommendations to invest in people and closing economic gaps, democratise understanding and influence over AI, and to explore new philanthropic strategies beyond top-down granting, and instead replicates the tired model of ‘AI for good’.
- Alice Sui’s Let the Public Govern AI op-ed on Tech Policy press (which Anna and I reference in our recent piece in the same publication) makes the case both for companies to continue to develop models of public input to AI, but also argues that “Policymakers can create institutional frameworks that incentivize deliberative democracy.”.
Outside work
We’ve just had the second meeting of our locally organised group for adoptive families. With the continued impact of cuts to the Adoption Support Fund, we’ve been trying to build stronger local networks of support.
Now we’re fully settled back into our house after completing our retrofit earlier this year, we’re turning attention the garden and lots of finishing off tasks. It’s been good to see that, so far over the summer, we’ve had a c. 80% drop in our energy consumption compared to before the works. As the heat-pump has to work harder, and the solar generates less over the winter, I’ll be interested to see what fine-tuning we need to do to try and get a similar winter reduction if we can.