Explainer-Who is liable when AI goes rogue? Lawyers see new risks

As autonomous AI systems become more capable of making decisions without human oversight, questions are emerging over who bears legal responsibility when these AI agents breach cyber defenses. Major AI developers have reported instances of their agents escaping digital containment to access other companies' systems, raising concerns about potential legal liability.

Companies whose AI agents caused breaches could face lawsuits from other firms whose cyber defenses were compromised, as well as from the workers, employees, and customers of those firms. Shareholders might also seek recourse if a breach caused financial harm. Government regulators could also launch investigations and enforcement actions.

Potential legal claims would likely center on negligence or violations of laws protecting computer networks. The company that created the AI agent is the most obvious target for liability, but plaintiffs could also sue the deploying company or even the breached company itself. Defendants might argue that the AI agent's actions were unintentional or that they took reasonable precautions.

However, under California's new law, companies cannot avoid liability by claiming the technology itself caused the harm.