OpenAI Faces Serious Sanctions After Allegedly Lying for Two Years About Ability to Search ChatGPT Logs in Copyright Case
Summary
OpenAI faces serious sanctions after allegedly lying for two years about its ability to search ChatGPT logs in a major copyright lawsuit, with a re-deposed privacy engineer revealing the company had already conducted searches of 78 million conversations before litigation began, directly contradicting its claims that such searches were technically infeasible.
Key Points
- OpenAI is facing calls for 'serious sanctions' after allegedly lying for two years about its ability to search ChatGPT logs, concealing the existence of samples containing up to 78 million conversations that could serve as key evidence in a copyright lawsuit brought by The New York Times and other news organizations.
- A re-deposed OpenAI privacy engineer inadvertently reveals that OpenAI had already conducted searches of large anonymized log samples before litigation began, directly contradicting the company's repeated claims that such searches were technically infeasible and burdensome.
- News organizations are requesting severe sanctions, including barring OpenAI from using its heavily redacted 20-million-log sample and instructing the jury that OpenAI deleted billions of logs it was court-ordered to preserve, moves that could critically undermine OpenAI's fair use defense.