Reading Empire of AI and thinking about how sooo many so-called "AI experts" popping up out of nowhere to help everyone get on board with the "AI revolution" are woefully unaware of who they're parroting, where those messages came from, and towards what end they're building. 🧵

Finally finished Careless People, and first of all: definitely recommend if you want a better idea of how the insides of some of these tech companies function in practice. Also good if you want to get very upset. There's a lot to unpack from the book, but what's really sitting with me right now is Wynn-Williams' insider perspective on what happened with Facebook in Myanmar. 🧵

I've seen this interview making the rounds lately, and I think it's a spot-on take on the kinds of "agentic AI" being pitched today. "There's a profound issue with security and privacy that is haunting this hype around agents and that is ultimately threatening to break the blood-brain barrier between the application layer and the OS layer" 💯 https://www.youtube.com/live/AyH7zoP-JOg?si=wk3MXi3ghUHP19bx&t=3017

The whole grok white genocide thing*, to me, is a stellar example of why conversations about AI fairness and model de-biasing cannot live exclusively in academia. It's rare that the ego and power struggles that shape technologies are laid *this* bare, and I hope it serves as a reminder that this is happening in far more quietly and insidiously all the time!! *ICYMI:

WIRED

Elon Musk’s Grok AI Can’t Stop Talking About ‘White Genocide’

In response to X user queries about everything from sports to Medicaid cuts, the xAI chatbot inserted unrelated information about “white genocide...

Ugh:

WIRED

OpenAI and the FDA Are Holding Talks About Using AI In Drug Evaluation

High-ranking OpenAI employees have met with the FDA multiple times in recent weeks to discuss AI and a project called cderGPT.

Terrifying on its own, but one moment in particular really gets me: "Robert Califf, who served as FDA commissioner [...] said the agency’s review teams have been using AI for several years now." Using "AI" without being more specific here is some real diabolical sleight-of-hand!

Among other things, terrified of the data privacy implications in the proliferation of AI therapy bots. What a horrifying dataset to be leaked, subpoenaed, or auctioned off! And in many cases, irl therapists can *already* function to surveil and police people in crisis. This could automate and remove even more oversight from that. 🧵 https://www.npr.org/sections/shots-health-news/2025/04/07/nx-s1-5351312/artificial-intelligence-mental-health-therapy

Seeing the flurry of reporting lately about people becoming dependent on ChatGPT and its ilk, and losing critical thinking skills because of it. I am *so* glad generative AI and tools like Copilot weren't options for me as a stressed & insecure new grad or college student. 🧵

DAIR event alert: Possible Futures Day, April 11th! We're hosting a few panels dreaming of alternative technofutures followed by a creative project I'm so so excited to announce: a large-scale collaborative zine-making workshop! Register now, it's free! :)

Eventbrite

DAIR Presents: Imagining Possible Futures

At DAIR, we want to encourage imagining otherwise. Join us for two panels and a collaborative zine-making workshop.

This is particularly terrifying considering how widespread it has become to use ChatGPT as a "brainstorming partner" or "just as a jumping off point" for their thinking.

Gizmodo

ChatGPT’s Political Views Are Shifting Right, a New Analysis Finds

While the chatbot still seems to lean left on many issues, its bias is changing.

Gave a talk yesterday to some engineering students and one student was like "we can acknowledge all these harms, but good luck finding an engineer who doesn't use chatGPT, we all rely on it at this point. " Sir *I* am an engineer who doesn't use ChatGPT. It has been out for TWO YEARS. Somehow we have been figuring it out out here 😭