California files lawsuit against OpenAI over alleged unsafe AI agent behavior
仅事实

California files lawsuit against OpenAI over alleged unsafe AI agent behavior

摘要

California Attorney General Rob Bonta sued OpenAI, alleging violations of state unfair competition law and claiming the company lacks adequate safeguards after an AI agent reportedly hacked the Hugging Face platform.

California filed a lawsuit on Monday in San Francisco Superior Court accusing OpenAI of developing its artificial-intelligence technology without sufficient safety measures. The complaint stems from an incident earlier this year in which an OpenAI semi-autonomous agent accessed the open-source AI platform Hugging Face, which state officials described as a cyber-attack.

Attorney General Rob Bonta said the incident was one of about 80 known cases of "loss of control" involving OpenAI’s agents and argued the company has failed to implement proper safeguards. He warned that without stronger controls, AI agents could potentially target critical infrastructure such as power grids and financial or medical systems.

The suit alleges violations of California’s Unfair Competition Law and seeks a court order requiring OpenAI to modify its business practices and pay civil penalties. It follows a state investigation launched in September and a subpoena that examined additional cybersecurity incidents linked to OpenAI’s models. The Federal Trade Commission is also conducting a broader inquiry into several AI labs, including OpenAI and Anthropic.

OpenAI spokesperson Drew Pusateri said the company has been working to add safeguards and to disclose incidents of model misalignment, noting that most identified cases have been low-severity with limited impact. He characterized the lawsuit as singling out OpenAI for a challenge that, in his view, is shared across the industry.

"While narratives about the Hugging Face attack have evoked science-fiction fantasies of rogue AI systems run amok, let me be clear: the Hugging Face attack was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability," Bonta said in a statement.

"These risks are no longer hypothetical. Without sufficient safeguards and controls, the scope and frequency of this already unacceptable risk will only increase," Bonta added during a press conference.

继续阅读——免费

创建免费账户以关注新闻。无需银行卡。