cab

Wow. It really is end times - now a Gail’s outside Brixton tube. Jesus, this is a real impetus to get out of London.

Just helping with editing here. These bots aren’t sentient. They’re doing what they’re programmed to do. OpenAI, a private company supposedly bound by laws, committed federal crimes.

Folks, if you want to know what's *really* up with "AI", where the harms lie, and contextualize the nonsense du jour, this film is must watch.

So they prompted their model running without guardrails to hack some shit. OpenAI told it to do that, the model didn't do shit autonomously (no LLM ever does anything autonomously, it's always prompted).

Shock horror. ‘AI’ summaries don’t in fact summarise. It’s so frustrating to read reporting. As if the default position is to assume the parlour trick will do anything the marketing says. This isn’t what these toys are made to do. www.the-guardian.com/business/202...

"I am part of multiple rapidly expanding global networks of people & communities who are resisting AI & objecting to the power of the big tech companies. Within these resistance networks, women are leading the way" a must-read from my colleague Jennie Stephens www.irishexaminer.com/opinion/comm...

"In April, the Metropolitan police launched investigations into hundreds of officers after using an AI tool built by the US tech company Palantir to root out rogue officers." #privacy #surveillance www.theguardian.com/technology/2...

The AI consciousness debate is a trap, argues Randima Fernando, co-founder of the Center for Humane Technology. The real question is not whether or not a model “feels” fear when its internal workings light up; it is who profits from blurring that line and who gets harmed.

AI isn't a thinking tool that occasionally imagines things which are not correct as an error/failure mode. It's a tool for generating mock text that looks very realistic(and realistic implies plausibly related to real world phenomena and organisations) without doing the cognitive work.

So much wrong with this its impossible to know where to start. 🤦🏼‍♀️ www.theguardian.com/science/2026...

Mainstream media discovers TESCREALism without crediting the people who did the defining work. @timnitgebru.blacksky.app @xriskology.bsky.social www.theguardian.com/us-news/ng-i...

if you are "using AI for writing," you are not writing. i do not know how to make this any simpler

One of the things i keep trying to get people to understand is that these linguistic shorthands and stylistic flourishes have existed in writte and spoken communication for long before "AI" got ahold of them, and their proliferation with the training data is a part of where "AI" "learned" them. …

"Rubbing stuff on your skin may have its flaws, but specific substances have proven themselves to be groundbreaking --- if not lifesaving --- like sunscreen."

So we’ve been cooking up some thoughts on Gen AI at @amnesty.org and I’m ecstatic to finally see it out in the world! Our 🗝️ conclusion: standalone gen AI systems, built on unlawful web scraping, constitute mass invasions of privacy by design & should be prohibited www.amnesty.org/en/latest/ne...

Not really an audio person but think linguistics is neat? All Lingthusiasm episodes have high-quality, human-made transcripts, with proper nouns and examples in other languages spelled correctly! (Okay, no guarantees the typo gremlins have NEVER crept in but if they have they've done so honestly)

“Why are women less bullish overall on AI than men” is not actually that mysterious of a question

I can think of no better act of resistance to #language oppression. Is it any wonder that speakers continue to see Creoles as ‘broken French, English etc’ when their very own elected representatives continue to describe it as such.

If you are not reading Emily Bender on this subject, you should be. She’s an essential voice.

So pleasing and heartwarming, especially the cheering

Jeez, things are not great inside Google www.linkedin.com/posts/amshal...

now thats an anecdote with a punch, from Gillian Tett www.ft.com/content/c0ae...

Saying this partly as I'm seeing lots of Labour mentions of deliverism and fixing state capacity pop up this morning, but another idea would be for ministers to stop pushing so many silly tech policies and also assuming AI will magically fix everything without anyone attending to the basics

inspired by a post I saw on here and have since lost, I've started reading "AI companies" as "cigarette manufacturers," so something like this becomes "maker of Newports backs bill to fund smoking literacy in schools"

In totally unsurprising news, Richard Dawkins is developing AI psychosis. Paywall bypass: archive.is/6RdK9

All of this thread is so important. I’m in that stage of hormonal deficiency where formulating thoughts is difficult but reading the article this morning made me rage.

A very thoughtful review of Doctorow‘s "enshitification" by @uxmichael.co, especially the latter sections about the pasts of the internet. www.first-and-fifteenth.co/first-18-ens...

For a quick overview of the early history of Hansard (and some of its rivals), see this post from our editor Dr Philip Salmon: victoriancommons.wordpress.com/2017/11/17/r...

Read our new report in @thenerve.news. Palantir has been able to take root in the UK because lobbying & spinning is so easy here. We have a tiny journo-political class that incessantly overlaps, a hotbed for cronyism that’s enabled Palantir to infiltrate every aspect of state 5/

Ghost in the Machine (@ghostdoc2026.bsky.social) is a must-watch documentary, now showing in London @berthadochouse.bsky.social dochouse.org/event/ghost-... With @emilymbender.bsky.social , @adambecker.bsky.social & many smart people documenting the dark roots of so-called ‘AI’. #notai #quitgpt

when people talk about AI used in moderation of social media posts, you must remember there is always the human-in-the-loop, often a woman, often in the Global South, always dehumanised and forgotten 1/n www.theguardian.com/global-devel...

#TalkAboutHumanities We need scholars across the humanities, because these are the fields where we study what it is to be human, to inhabit different identities, and to connect with each other, to be human together. youtu.be/T7Lc6QNxolQ...

We are super excited for the next Mystery AI Hype Theater 3000 when @alexhanna.bsky.social and I get to talk with Carmen Maria Machado about the use of LLMs in writing, from language arts classes to the publishing industry. Livestream info: Monday, April 13, noon PT, twitch.tv/dair_institute

irisvanrooijcogsci.com/2026/04/08/r...

"The use of ChatGPT for scientific purposes... clashes with core scientific integrity principles such as honesty, scrupulousness, transparency, responsibility, and independence. Therefore, I need to decline this invitation to review."

"It is not about whether AI will help itself to your job. It is about whether the people who make AI are helping themselves to your country."

I worked on a thing! (With @imaginaries.bsky.social and Jascha Bareis from the @hypestudies.bsky.social group.) akademie.dw.de/hype-literacy/

This is disappointingly-credulous xrisk palace intrigue, utterly failing to contextualize the xrisk/AGI narratives. The public would be better served reading @timnitgebru.blacksky.app and @xriskology.bsky.social on TESCREAL or watching Ghost in the Machine notaidoc.com @ghostdoc2026.bsky.social

Pacific north west-on super mare

This headline is bullshit. Just more of the ‘I wrote ‘I am alive’ and put it in the copier and the result has shocking implications’ From the article: “[The bot owner] told me that Tom [the bot] wrote these blog posts, but that he “might have suggested” Tom write about these specific topics.”

AI newsletters sound more and more like when the STEM kid eats an edible. Political superintelligence would be so amaaaaazing. If it was technically possible and we could devote previously unavailable political coordination to reorganise all of society...

Yet more incredibly bad tech reporting from the Guardian. How many times? There is no ignoring. There is no scheming. There is no agency. This is synthetic text. For responsible tech journalism, subscribe to @404media.co www.theguardian.com/technology/2...

To the two insecure teen boys on the train this evening: I feel for you. Please don’t do ‘pinning’, steroids, ozempic or any of the other things you were talking about. Just be kids while you still can and ignore the online grifters. Ride your bikes fast, listen to music loud, laugh hard. Live.

I cannot recommend this enough! Was so excited to watch on Kinema this weekend and it didn’t disappoint. notaidoc.com