Today on @indicator.media: A first-of-its-kind audit of AI labels on major social platforms.

Could not doom-scroll earlier today? Yeah, me neither bc of our dangerous reliance on hyperscaler clouds

#AWS us-east-1 outage took down #Signal and chunks of the internet. The culprit was too many critical operations crammed into a single AWS data center cluster in Northern Virginia.

The use of AI generated content in political messaging has become a feature in today’s communication ecosystem.

How often is it used in the 🇳🇱 election campaign, and by whom?

Check out colleague @favstats.eu and @meinungsfuehrer.bsky.social’s AI campaign tracker #tk2025

www.campaigntracker.nl/en/

Thread - I've been thinking a lot about Walter Lippmann’s Public Opinion lately, especially in the context of Musk’s role in dismantling the federal workforce through DOGE. It’s a perfect example of what Lippmann warned about over a century ago.

There's obviously not enough to worry about right now, so I thought, hey why not drop an episode about deepfakes...

Fortunately, what @samgregory.bsky.social told me did make feel better. We discuss what technologists can do to help protect our shared reality:

open.spotify.com/episode/6HE6...

This is huge in terms of the fair use piece.

Mixed emotions from Paris today, but immensely enjoyed the rich conversations at the side events I had the chance to attend. Most focused on the power of deliberative democracy, trust and safety, the benefits of AI regulation, the potential of open source and themes of sustainability and inclusion.

Great discussion from @f-mb.bsky.social, Catherine Régis, @ginasue.bsky.social and Célia Zolynski on their recommended actions for AI and elections at the Sorbonne.

En route to Paris to join some of the discussions on AI safety happening at the AI Action Summit. Very happy to be returning to the City of Light almost a decade after studying at SciencesPo!

DeepSeek forces a rethink of the compute & energy costs of AI models. In a new paper with @sashamtl.bsky.social @strubell.bsky.social, we look at the full environmental impacts of AI – both direct and indirect – and what 𝐉e𝐯𝐨𝐧𝐬 𝐏𝐚𝐫𝐚𝐝𝐨𝐱 means for AI and climate. A thread 🧵

Visited de Koepel in Haarlem over the weekend, a former panopticon-style prison turned cultural space (slightly strange, with the questionable inclusion of an escape room...). Stark reminder of the role of surveillance practices and their extension into our daily lives.

Great & timely research that provides rich insights into safety issues with some open VLMs and across different languages 👇

thereader.mitpress.mit.edu/the-staggeri... It's not one to discuss scale but it's by far the text that's ignited most engagement on AI and environment among technical undergrads, that I've seen

Setting the scene for a head-on clash between Facebook and its EU regulators (where enforcement is split between Ireland/EU Commission).

Link to the paper: arxiv.org/abs/2406.17737

The number of risk assessments are coming in #DSA #commsky #eusky

Check this overview ⬇️

X, Facebook, Instagram, Google, YouTube, TikTok are there

docs.google.com/spreadsheets...

Had already stepped away from the noise of X/Twitter for quite some time, but on days like today, I'm seeking meaningful connection. So, hello Bluesky - here to listen, share & hopefully find thoughtful discourse.

Here's to brighter skies ahead...

In our latest investigation, Bellingcat researcher @koltai.bsky.social reveals the hidden network behind OpenDream, an AI art generator platform. She found several instances of child sexual abuse material being generated and featured on the site. www.bellingcat.com/news/2024/10...

From Toni Morrison's 1995 lecture "Racism and Fascism"