Skip to main content
All Stories By:

James Vincent

James Vincent

Senior Reporter

I mostly cover machines with brains for The Verge, despite being a human without one.

AI is killing the old web, and the new web struggles to be born

The capacity of AI to generate content is overwhelming the web

External Link
OpenAI reportedly trained its AI models on YouTube.

That’s according to a report from The Information on the value for Google of YouTube as an AI training dataset. The fact that OpenAI scraped YouTube isn’t surprising, but the company is famously secretive about its training data, partly for competition reasons, and partly, it’s thought, to stymie potential lawsuits.

YouTube’s terms of service forbid using content for anything other than “personal, non-commercial use,” but it’s an open secret in the AI industry that everyone is scraping the web constantly. If Google protests too much, it would end up incriminating itself.

External Link
Are you an information prepper?

The spread of AI-generated misinformation is worrying lots of us, including Ars Technica’s AI reporter Benj Edwards. That’s why he bought the only remaining and regularly updated physical encyclopedia in existence. As Edwards explains: it’s expensive, sure, but at least it’s reliable, and won’t suffer link rot or stealth edits. Is it time to stock up on facts while we can?

Stable Diffusion has a big bias problem.

Bloomberg has done an amazing deep dive into the racial and gender biases in popular AI image generator Stable Diffusion.

Sadly, the results are exactly what you’d expect: generated images of people in higher-paying jobs are disproportionately white and male. And this pattern doesn’t just reflect social biases but amplifies it, as the graphic below shows.

Check out the full story here.