How OpenAI got hacked with an image

Two individuals successfully exploited a one-year-old libheif heap overflow vulnerability to gain remote code execution on OpenAI's Discourse forum. This allowed them to compromise an employee's ChatGPT account and leave a message within the internal monorepo. This incident highlights how AI is altering the economics of exploit development and demonstrates the ineffectiveness of security through complexity in the current threat landscape.

This report uniquely details how a one-year-old vulnerability was leveraged to breach OpenAI's systems, unlike general discussions of AI security.