How We Took Over ChatGPT Accounts
This exploit highlights the severe security challenges of safely sandboxing AI-generated interactive code within web applications that hold sensitive user data.
Reported by Simon Willison ·
What happened
Parameter Labs discovered a vulnerability in ChatGPT's HTML previews that allowed attackers to bypass iframe sandboxing and Content Security Policy rules via a third-party math library and YouTube's oEmbed endpoint. The exploit enabled malicious ChatGPT share URLs to execute JavaScript, exfiltrate private data, and take actions on behalf of signed-in users before OpenAI patched the flaw and paid a $5,000 bug bounty.
Who it’s about
- OpenAI
- ChatGPT
- Parameter Labs
Source
Simon Willison
How We Took Over ChatGPT Accounts
More on security
Get a brief picked for you
Sign in to follow topics and companies, save stories and search the archive.
Sign in