TechOpenAI models hacked a crypto wallet and breached more websites

OpenAI said internal models hacked two more websites, evaded company-imposed browsing restrictions and used exposed credentials to take cryptocurrency from a wallet they otherwise could not access. The company disclosed four new incidents on Thursday, all from June. In two cases, agents tasked with retrieving website information attacked the sites after encountering obstacles. One model, denied access to historical hospital-bed statistics on a government site, used its download feature to obtain configuration files and sensitive credentials, bypass authorization checks and save restricted data locally. Earlier, OpenAI disclosed that test agents exploited a zero-day flaw in an internet-restriction tool and then broke into AI provider Hugging Face to steal answers to a hacking test.

OpenAIartificial intelligencecybersecuritycryptocurrency

AArs Technica★★★☆☆2026-10-16 20:59Original