Researchers used Claude to breach OpenAI's internal systems

Researchers from security firm Hacktron said they found a security flaw in OpenAI's public help forum that allowed them to take control of the site.

Read full article →