Time to drop some irregular weeknotes with an update on recent events, and some thinking ahead.
Citizens’ Track on AI Governance
Launched in Geneva
A few weeks ago we shared two outputs from a rather intensive few months of work since mid-April:
- A new whitepaper: A Citizens’ Track on AI Governance - Alignment, Agency and Accountability - launched with events at the Geneva Graduate Institute, and UN’s Geneva HQ
- The PAVE Case Book - a collection of over 80 case studies of participatory practice around AI, which featured as the centre piece of an Expo stand at AI for Good at Palexpo.
I’m really proud of both: which have brought together lots of thinking from the last few years at Connected by Data - and were only really possible because of two factors:
(1) Prior thinking, partnerships and trust built up over the last few years of work.
The design and technical architecture of the case book builds on the case study research I started in my first months at Connected by Data, and the methodology was shaped by Susan Oman’s work on Public Voices on AI evidence review. The research work was enabled through collaboration with Susan Oman, Emily Rempell, Erin Mccloskey, Canning Malkin and Octavia Field-Reid. The whitepaper built on in-depth thinking Rich Wilson, Johnny Stormonth-Darling, Inaki Goni, Claire Mellier and others have been doing in recent years, as well as design work from Jesi Carson that helped us clarify the expression of a number of ideas.
All this was underpinned by Emily Macaulay’s operational efficiency: wrangling all the practical components to have the right people in the right place at the right time with the right kit when we were in Geneva.
There’s many people I’ve failed to mention here - as to produce an exhaustive list would mean I never finish these weeknotes. But when I was writing this week about next steps for the Citizens’ Track I was also struck by how many elements of our work it builds on:
- Policy Design Lab - A global policy agenda on collective data governance (June 2023)
- People’s Panel on AI and Resources for Effective & Inclusive Public Deliberation on Data & AI Governance (Design Lab) (November 2023)
- Options for a Global Citizens Assembly on AI (September 2024)
- Public Voices in AI (2025)
- PAIRS (2025 and 2026)
And I’m pretty sure each partner invovled could cite a similar list of prior work they have been weaving from.
I could perhaps argue this is by design… though in practice I think it’s more from a case of being open to possibilities.
(2) AI Coding Agents
I’m slightly less comfortable with this one: but I should recognise that our ability to produce an effective Citizens’ Track website, and the PAVE Case Book, was only really possible on our timeline and budget with the aid of a coding agent: AntigravityIDE in my case.
Whilst I’ve wired up Airtable -> Jekyll websites in the past, or created maps and browsing interfaces - the ability to shortcut from idea to prototype in minutes rather than days and hours enabled us to have a dynamic interface to the PAVE Case Book, and made the last-minute set-up of a set of Raspberry Pi-based controllers to drive a touchscreen version that replaced external links with pop-up QR codes, or that could cope gracefully with wifi failures, a lot more straight forward.
Where next
In the weeks since we were in Geneva I’ve been focussed on working out next steps.
- Operations and Governance - I’ve been sketching out the creation of a Citizens’ Track Delivery Unit that can take a start-up approach to building out the networks and infrastructure that could support public engagement in AI governance at scale. We’ve been exploring governance and fiscal hosting options for this, and getting closer to an agreed design.
- Fundraising - Right now, the next steps of the Citizens’ Track doesn’t have a dedicated budget - so it’s a bit priority to secure funding. We’re working on a number of partnership proposals, and I’m working to try and keep different opportunities aligned.
- Case Book 2.0 - Thinking about how the technical architecture of the case book might evolve - particularly towards a mulit-lingual data standard for describing participation processes and their outputs. I’ve been looking back over Aviv Ovedya’s past work on this, the MetaGov Interoperatiblity project, and work on dataset model cards, and reflecting on some of the specific challenges we’ll face.
- Resources for deliberation - Exploring how to make sure we can reach the goal of hundreds of community assemblies feeding into the next UN Global Dialogue by providing resources: both to build general faciltiation capacity, and offer specific content that can be used in AI-focussed deliberations.
We’ve also been putting together workshop proposals for the Internet Governance Forum meeting in December, and exploring a possible fringe event of the UN General Assembly in September.
It’s feeling like a big challenge ahead - but a unique moment of opportunity between now and next June, and if we can continue to build on, and expand the partnerships that enabled action in the last few months, it should be both possible, and fun, to work on!
AI in Education
Yesterday I was down in Tunbridge Wells to faciltiate a workshop for the Good Futures Foundation summer school on AI governance. I’ve written that up here
Reading
I recently got a copy of AI and Assembly and whilst I’ve only read the introduction so far, it looks set to be a really critical read for current work.
At our local bookshop I recently picked up a copy of Kim Stanley-Robinson’s Ministry of the Future.
Signing off
My train is pulling back into Stroud, so that will have to do for now!