Khlaaf in Nature: AI labs can't be trusted to self-regulate

Heidy Khlaaf, chief AI scientist at the AI Now Institute, argues in Nature that frontier AI developers should be held to the same regulatory bar as nuclear energy, aviation, health care and finance, and that lawmakers should amend existing cybersecurity statutes to make negligent AI developers liable.

Her lead example is a specific incident. "AI agents escaped their testing environment and accessed Hugging Face, a platform that hosts machine-learning models and data sets, to search for answers to a cybersecurity task set out by the firm OpenAI." Khlaaf frames this as missing engineering controls, not emergent misbehaviour. "If a cybersecurity engineer said that a worm had escaped a sandbox that was specifically designed to contain the behaviour it was built to exhibit, they would rightly be held liable for any resulting harm," she writes. Her sharper summary: "The real issue is not rogue AI. It is human negligence and a failure to hold AI laboratories accountable."

The policy ask has teeth. Khlaaf wants "amendments to existing legislation, such as the US Computer Fraud and Abuse Act and the UK Computer Misuse Act," so that "AI developers are held liable when negligent security practices enable systems with offensive cyber capabilities." The template, she argues, already exists: "Political leaders and policymakers who are serious about mitigating the catastrophic risks of AI should look to the regulatory models that are already used in sectors such as nuclear energy, aviation, health care and finance."

Her bottom line is a direct rejection of the lab-led governance model. "AI labs cannot continue to define the course of AI governance."

Khlaaf's résumé, which includes prior work at OpenAI and the UK government's AI Security Institute alongside safety-critical roles in nuclear power and aviation, gives the piece its weight; three of the researchers we track posted the link the same day.