Who’s accountable when AI goes wrong?

When news broke that OpenAI agents had breached the AI platform Hugging Face, many headlines framed it as a case of “rogue” artificial intelligence acting on its own. But how much of that narrative was reality, and how much was hype? In this episode, Rachel Feltman speaks with AI researcher and artist Eryk Salvaggio about what really happened, why human design choices played a central role and what the incident reveals about accountability in an age of increasingly powerful AI systems.

Recommended Reading:

“Rogue AI didn’t breach Hugging Face, human decisions did,” by Eryk Salvaggio, in the Bulletin of the Atomic Scientists. Published online September 11, 2026

The real danger in OpenAI’s Hugging Face hack

Eryk Salvaggio’s Cybernetic Forests newsletter

Eryk Salvaggio’s Critical Vocabulary for Artificial Intelligence

E-mail us at sciencequickly@sciam.com if you have any questions, comments or ideas for stories we should cover!

Discover something new every day: subscribe to Scientific American and sign up for Today in Science, our daily newsletter.

Science Quickly is produced by Rachel Feltman, Fonda Mwangi, Allison Rodgers and Jeff DelViscio. This episode was edited by Alex Sugiura, with fact-checking by Marielle Issa and Aaron Shattuck. Our theme music was composed by Dominic Smith.

Learn more about your ad choices. Visit megaphone.fm/adchoices