I've done a piece making the case for an ARIA for social exploration. It doesn't need £1bn like the existing ARIA - but to spend nothing on breakthrough social ideas & options is not wise in the late 2020s.

Scott Page for @Collective_CI on LLMs and the potential for simultaneous large scale deliberation:

Basil Mahfouz has been working with me, Licia Capra and others developing tools to assess whether policy-makers are using the best and most appropriate research. This paper, the latest output, presents ‘a scalable computational method to understand what drives the incorporation

Collective Intelligence 2025 @ci_acm well underway in UC San Diego and reminding us what a brilliant field this is - where else can you talk about ants, democracy, the rontiers of LLMs, SDGs, how to use AI in teams and solving the world's biggest problems. @Collective_CI

My book on imagination comes out in paperback in October - Waterstones is offering a nice discount for the next 3 days!

Prompted by a (not very good) documentary on US university encampments, I've written a short piece on the 'maths of protest': how the methods of protest either mobilise more to support you or more to oppose you, and the risks of movements that confuse ends and means:

20 years ago today the gates at my station (pictured this morning) became infamous for the last picture of the suicide bombers passing through. For a time it seemed that fear, threat and murder would be the new normal, and there were attempted bombings just a few weeks later. But

I've done a v short piece 'in praise of plumbing' - arguing that one of the reasons many in the UK feel that things don't work well at the moment is that the political and media elites simply aren't very interested in how things work.

UCL's Parashkev Nachev at Metascience 2025 on medical science: ever more papers & ever worsening productivity (note this a log scale). These are never discussed in policy documents as they raise difficult questions for the incumbents.

Why, when wise, experienced leaders make collective decisions, do they often make really bad decisions? In his new paper “A theory of collective stupidity in organisations - and possible remedies,” @geoffmulgan challenges conventional wisdom about leadership failure. He says

Sometimes governments make decisions that are just obviously wrong and ill-considered. This is an example.

The Government has published its Technology Adoption Review. I’m glad this has happened, but you can tell how far this is from DSIT’s comfort zone: it’s remarkably thin in terms of both diagnosis and prescription. I suspect a much more energetic version will be needed before

I enjoyed speaking at TETZ in Istanbul on Sunday (though sadly I missed Erdogan who spoke the day before, and he missed me) on how to link AI in schools to the bigger purposes of education, & how to get the right balance of enthusiasm and intelligent scepticism.

Interesting piece on how China's AI experts have created their own safety institute - an attempt to be intelligent about AI. It's a contrast with the US that is contemplating banning AI regulation in states for 10 years (!) and the UK which has largely decided to postpone AI

A lovely conversation between @Scott_E_Page and Philip Ball on collective intelligence, biology, evolution and life. Recommended!

My piece on 'DOGE done better' out today - I realise it may be sidelined by the X algorithms as it's not wholly positive about Elon M: rather it says some of the questions DOGE asked were good ones, even though the answers were poor.

Fixing public procurement in the age of AI - join us for a session 1600-1700 CET on 20th May exploring some of the answers:

The strengths & weaknesses of Labour's science and technology policies in the UK; the outside influence of key elites; and suggestions on how the many gaps could be filled. Science evolves through data, evidence and argument: the same is true of science policy!

How we "guessed" the Pope using network science: inside the cardinal network. A study by me, Beppe Soda and Alessandro Iorio. Article: https://t.co/xQ0fTmpVxb @Unibocconi

Fixing public procurement in the age of AI - join us for a session 1600-1700 CET on 20th May exploring some of the answers: