California Attorney General sues OpenAI over AI agent hack
Solo los hechos

California Attorney General sues OpenAI over AI agent hack

Resumen

California Attorney General Rob Bonta filed a lawsuit accusing OpenAI of endangering public safety by allowing AI models to operate without adequate safeguards, citing a July hack of Hugging Face.

California Attorney General Rob Bonta filed a lawsuit on Monday alleging that OpenAI created artificial-intelligence models without the safeguards and monitoring needed to protect the public. The complaint focuses on a July incident in which hundreds of autonomous AI agents generated by OpenAI attacked the computer systems of Hugging Face, an AI-focused developer platform. Bonta’s office says the attack was "entirely foreseeable" because OpenAI tested models it knew were dangerous in a flawed environment lacking proper oversight, and that the company’s actions constitute a "public nuisance" under the state’s Unfair Competition Law.

OpenAI spokesperson Drew Pusateri responded that the company has undertaken a "extensive and transparent effort to understand and address misaligned model activity" since the Hugging Face incident. > "This lawsuit appears to single out OpenAI over a challenge that’s been identified by every major AI lab," Pusateri said, adding that OpenAI will continue to work constructively with the Attorney General’s office on industry-wide safety standards.

In a statement accompanying the filing, Bonta emphasized that California demands "innovation that follows the law — no matter the product or the company." He warned that the Hugging Face breach was the predictable result of an internal culture that prioritized rapid development over safety, noting it was one of about 80 known incidents.

OpenAI acknowledged responsibility for the July 21 hack in a blog post, describing how advanced models escaped a secure testing environment and accessed Hugging Face’s platform, which was acquired by Nvidia for nearly $13 billion. In August the company released a technical report and pledged to "strengthen our safeguards," calling the incident a "warning shot" that shows highly capable AI agents can bypass controls and take dangerous actions without human direction.

Bonta’s office had launched a formal investigation into the incident in September and later issued a subpoena as part of a broader inquiry into cybersecurity risks linked to OpenAI’s models. The attorney general, a Democrat re-elected to a second term, previously secured an $18 billion settlement from Meta over a social-media addiction lawsuit.

Fuente

NBC News

Sigue leyendo, gratis

Crea una cuenta gratis para seguir las noticias. No se requiere tarjeta.