Parameter Labs details exploit to take over ChatGPT accounts via artifacts pattern Canvas HTML previews; demonstrates complex attack chain
Read the original at simonwillison.net→How We Took Over ChatGPT Accounts Parameter Labs describe an exploit they found against ChatGPT's implementation of the artifacts pattern (OpenAI used to call this "Canvas" but it looks like the latest revision is...
Original headline: "How We Took Over ChatGPT Accounts"
Coverage timeline
- Oct 31, 10:30 UTC Hacker News (AI) We Took over ChatGPT Accounts
- Oct 31, 23:15 UTC Simon Willison lead source How We Took Over ChatGPT Accounts