AI agents are being sold as democratic and participatory: a way to decentralise AI by giving every citizen a personal proxy. But is replacing friction with proxies necessarily more democratic? Or does it shift political voice into infrastructure that remains standardised and centrally structured?

One big, cruel grift bankrolled by some the worst men - and a reminder that all education is political.

There will be lots of detailed analysis of this I'm sure, but a few immediate thoughts - it's premised on the absolute inevitability of more people using more AI at work, despite also saying that businesses are struggling to find viable use cases

That they chose “Lunch atop a Skyscraper,” a photograph of workers during the Great Depression, and superimposed the images of some of the top figures of the new Gilded Age tells you a lot about the role of labor in AI—it’s the Mechanical Turk all over.

Our article on the politics of AI prompts is out now @econsocjournal.bsky.social www.tandfonline.com/doi/full/10.... 1/n

"Anatomy of an AI Coup" examines a shift of power from policy makers and bureaucrats to those with the skill to train and fine-tune large language models - often behind system prompts, which further calibrate the gov. worker's options toward the decisions of those who calibrate the system.

Quite aside from anything else, this is what it looks like in Gaza as people try to return home. (📷: Al-Jazeera)

India’s digital public infrastructure is going global—open-source, low-cost, scalable, writes Anuradha Sajjanhar. But big questions remain, she says: What kind of states are we building? Whose power does it serve? What politics are other nations importing? www.techpolicy.press/indias-digit...

We should have asked these same big questions in regards to US digital infrastructure...

I write about India Stack becoming a key national export and ask a question that goes beyond the region: As digital governance is scaled, exported, and repackaged as a global good, we must ask not just what it can do, but also what it displaces.

Here is a gift link to @kashhill.bsky.social's must read piece on the dangers of OpenAI's sycophantic LLMs and the parasocial relationships people are creating with them. www.nytimes.com/2025/06/13/t...

I don’t know if it’s a masculinity thing or a managerial class thing or what. But you don’t actually have to hand it to AI. It’s okay to look at the equation and conclude it’s not worth it.

Horrible www.technologyreview.com/2025/05/20/1...

Not only can it not help me with my research or writing, but its surreptitious adoption by students makes teaching significantly harder, as half the job is now working on the assumption that a range of standard assignments no longer teach human beings anything, because they can simply hand them off.

I keep catching strays from "technology is neutral" people, so let's set things straight. 1) Technology is not neutral, especially the subset of technology called "algorithms" 2) The form-factor that makes a technology into a tool is not neutral 3) How tools are productized is not neutral [rant]

📣UEA UCU members have voted for 9 days of strike action to commence this week from Thurs 1st May. 📣This is in response to management's plans to continue progressing compulsory redundancies. 📣There is still time for the employer to change course. #SaveUEA

Picketing with @uea-ucu.bsky.social with fierce toddler representation

Trump got in power & went straight after a) people & institutions that produce information & knowledge, and b) people & institutions that monitor & prevent fraud. I know the world is full of glassy-eyed poorly-informed dipshits, but honestly, how can you construct a benign explanation of this?

"Utterly delusional" does not even come close. archive.ph/OVHMg

'generative AI is experimental'

So one of the things that I think is lost on AI proponents is what I call the card catalog effect, a thing I shouldn’t call it because a lot of people probably have no experience with a card catalog.

Stand by this: www.politico.com/newsletters/...

I'm reminded of Lucy Suchman's excellent 2023 piece on "The uncontroversial ‘thingness’ of AI". Suchman urges us, in critical work on "AI" not to cede the ground that "AI" is a thing at all. journals.sagepub.com/doi/full/10....

The UK joined the US in not signing the international AI declaration pledging an open, inclusive and ethical approach to Artificial Intelligence. Earlier today @rachelcoldicutt.bsky.social argued that this may seem to be in the UK's interest, but in fact it isn't. blogs.lse.ac.uk/politicsandp...

This segue is doing more work than I could ever imagine.

The number of people surprised by the long list of words NSF is targeting makes me think many of you all have not seen this: www.commerce.senate.gov/services/fil...

"The 'doomsday trap' of artificial intelligence is the dehumanization that [it] makes possible. AI is not only a technology, it's...a spectacle designed as pretext to resist empathy and create emotional distance from consequences." @eryk.bsky.social mail.cyberneticforests.com/a-fork-in-th...

Much more reassuring press release from DSIT on AI in govt today, including some v welcome diligence measures www.gov.uk/government/n...

JFC. Musk, Bezos and Zuckerberg are in the front row at the #inauguration. They are sat IN FRONT of Trump’s Cabinet. The era of oligarchy. (📸: Julia Demaree Nikhinson, AFP)

Technology adoption is a social process, speeding it up requires social interventions backed up by useful infra; saying "open sesame" won't do it alone

Accurate

What if learning became effortless, guided by an AI tutor far superior to any human? If we no longer need a lengthy process of granular accumulation that ends in a mountain of expertise, will we have truly learned? newsletter.dancohen.org/archive/the-...

I wanted to say "i didn't need that, first thing this morning," but honestly, I'm going to need exactly this kind of motivation to get some shit done today

I always come back to this clarifying speech by @emilymbender.bsky.social on the term AI as "a way to make certain kinds of automation sound sophisticated, powerful, or magical...by making the machines sound like autonomous thinking entities rather than tools" medium.com/@emilymenonb...

The problem for me is that AI has become parlayed out of being "things computers can do that can be helpfully used to solve problems" into being a magical force that do everything, make everything better, replace everything, make all the money, and it's getting a bit embarrassing

"I never want to pretend that I’m inventing something new or starting some new conversation... It’s an exercise in intellectual humility to say: We as readers, I as a writer, am entering smthg that is ongoing. Novelty is fetishized in academia. Everyone wants to do the new thing." - Ruha Benjamin

This is useful:

My colleague @anuradhas.bsky.social on Starmer's AI pitch/pivot. She wrote a very fine book about Modi's government - how experts & think-tanks reshaped the BJP's/India's political & policy ideas. She knows what she's talking about. www.techpolicy.press/the-uks-big-...

I write on @techpolicypress.bsky.social, PM suggests that AI will make public services “feel more human,” giving us the “precious gift of time.” But such claims are strikingly devoid of detail. How will AI achieve these sweeping outcomes? What are the trade-offs? www.techpolicy.press/the-uks-big-...

Framed as a transformative vision for economic growth, national renewal, and cutting-edge public services, the UK announced its AI Opportunities Action Plan on Monday. The future may indeed be built by the AI industry, writes Anuradha Sajjanhar, but who ensures it’s one worth living in for everyone?

👋