Sam Altman
Role at the time: OpenAI chief executive, in an interview on POLITICO's Decoded podcast
Statement summary
Asked whether OpenAI can be held legally responsible under existing civil law for the Hugging Face breach or other breaches committed autonomously by its models, OpenAI CEO Sam Altman said he did not know enough about existing law to comment. He said there will need to be a liability framework for companies offering such models, possibly a separate one for AI training, and that OpenAI will need to be responsible for some version of what goes wrong during training.
“I do think there's going to need to be a liability framework for companies that are offering models like this.”
Why we used this label
The case asks whether OpenAI should face legal consequences for its agent's breach of Australian government systems and its delayed disclosure. Altman accepts that OpenAI should bear some responsibility under a future liability framework, possibly one specific to AI training, but declines to say whether existing law makes OpenAI liable and leaves the details to policymakers, so we labeled it Mixed or conditional. The question named the Hugging Face breach and other autonomous breaches, which includes the Australian incidents.