In the US or Canada, you can order the paper version here: cup.columbia.edu/book/the-ai-... End of month treat: on 30 November everything on the Columbia UP site is half price, including our book 😎

6️⃣ AI debates suffer from serious oversimplification: “AI” is not all the same. Generative AI is very different from computer vision, and AI-powered robotics differs from weather prediction. The PE effects differ, too – not all have the “winner takes all”-tendencies that we observe in GenAI.

5️⃣ Everyone talks about the US, China, maybe Europe and East Asia. Where are the other 6 billion people on the planet in conversations about AI politics? They, too, are implicated in what’s happening – whether as miners of essential minerals, remote click-workers or objects of AI imperial ambitions.

4️⃣ AI’s main effect on labour markets – at least for now – may not be that it’s killing all jobs. But that it’s dumbing them down, creating alienation and frustration.

3️⃣ AI may be good for GDP headline figures – but that doesn’t mean it makes most, or even many, people better off.

2️⃣ Even once AI has been developed, what happens next depends on political choices and institutions. “The AI transformation” can and should be politically shaped and steered.

A few core arguments:

1️⃣ From the get-go, governments and their political priorities have been implicated in how, where, and to what ends tech is developed. AI has been no different.

In the meantime, some things I found interesting:
@dmugge.bsky.social writes about lessons from Trump's win for EU's digital transformation: "If Europe wants digitalisation to bring genuine societal progress, the lodestar should not be a competitive race with the US" www.euractiv.com/section/tech...