Twitter Beğeni Hilesi

How did Annie Gilbertson prove that tech giants were using scraped YouTube videos in their AI models? She breaks it down for me here in a new series on Proof News where we dissect how the news is made: youtube.com/watch?v=XZkOcj…

I'm glad to see our reporting on how Meta has cut back on transparency being cited in Senate letter to Meta demanding greater researcher access. proofnews.org/meta-is-gettin… coons.senate.gov/imo/media/doc/…

With Google reneging on its promise to block third party cookies, are we in for even more cookie consent pop-up boxes — and less privacy? My discussion on NPR Weekend Edition: npr.org/2024/07/27/nx-…

Here's what happens when we put our AI testing tools in the hands of creators: They ask questions of the AIs and are shocked at the range of quality of the answers between models. This is how we get AI to be better: by getting them to compete on quality. 👇🏻

One weird norm in investigative journalism is that if someone scoops you, your story is dead. But it doesn’t have to be that way. That’s why we are building on the work of our friends at 404Media with a dive into how Nvidia scraped YouTube videos for its AI models

Great find by Jason Kint. Love that this Google expansion of online tracking was code named Narnia2 — given that it did actually transport every Internet user through a magical portal into a whole new level of surveillance.

So fun to talk about with Taylor Owen about how journalists can improve coverage of topics like AI 👇🏻

In the last 6 months, news organizations like Vox, the WSJ, and The Atlantic have signed content-licensing deals with #OpenAI. To Julia Angwin, this is a devil's bargain. Are journalists becoming content farmers for AI? Listen: tgam.ca/angwin

In the latest  Proof Ingredients video, @juliaangwin sits down with Carl Brown, creator of @InternetOfBugs, to discuss if AI could replace software developers. proofnews.org/is-ai-going-to…

This platform is an unlikely role model for governance. But I think Hank Green makes a fair point in asking YouTube to (at the very least) emulate X by offering creators an opt-out of AI training. m.youtube.com/watch?v=JiMXb2…

"One of the things that upsets me the most about AI is the myth that it will spit out the answer. The idea that there’s just one answer is a fiction—and a dangerous fiction,” says tech journalist Julia Angwin on Machines Like Us Podcast. Check out the full video: tgam.ca/3YZCYhQ

Building a better Web populated by higher quality pages is a goal we should all shoot for. Forcing Google to open up its data is key to getting us there. My latest for New York Times (gift link): nytimes.com/2024/08/27/opi…

I continue to be disappointed by the poor accuracy rates of AI. Our latest Proof News investigation shows the leading AI models were incorrect about one-third of the time when confronted with known misinformation about the presidential candidates. 👇🏻

Just announced: 3 amazing keynote speakers for C+J 👏🏽 Julia Angwin, Aaron Sankin, Alan Mislove. Join us Oct 25-27 in Boston at Northeastern. Registration live: idi.provost.northeastern.edu/event-director…