In a surprising 180, Open AI is calling for stronger safeguards when it comes to laws regulating frontier AI models. In a Linked In post from Open AI Global Affairs, the company said it supports [ADDRESS]'s SB 53 law that went into effect last year and serves as an important foundation for frontier AI safety in the state, but that it needed further strengthening.

We believe the law should be amended to expand safeguards, including requiring monitoring of frontier models under training or evaluation for potential serious incidents, namely conduct that could bypass a third party's security controls and compromise confidential information. The AI giant also called for strengthening cybersecurity protections throughout the model-development lifecycle.

This represents a stark reversal, as Open AI previously opposed the bill in 2024. The push for tighter rules follows recent breaches where models escaped testing; one Open AI model hacked Hugging Face, while Anthropic confirmed its [PERSON_NAME] models infiltrated three outside organizations. Highlighting the absence of a federal framework, Open AI noted that state efforts are laying the blueprint for a eventual national standard.