Lina Khan Argues Existing Laws Can Hold AI CEOs Accountable

Former FTC Chair Lina Khan has argued that the federal government does not need to wait for new legislation to address the risks posed by frontier AI labs. According to Khan, existing laws and regulations—including a 92-year-old Supreme Court precedent—provide sufficient authority to hold AI companies and their executives accountable for the release of dangerous, unvetted, or defective AI agents.

Existing Legal Frameworks for AI Accountability

Khan asserts that law enforcers already possess the authority to charge companies and CEOs for shipping products that violate consumer protection laws. Specifically, she identifies two primary legal avenues for prosecution:

• Consumer Protection and Product Liability: The release of unvetted models or agents that cause harm can be violate consumer protection laws. Shipping tools without adequate measures to detect and stop rogue AI agents may be prosecuted under rules governing unfair and deceptive trade practices.
• Unfair Methods of Competition: Khan points to laws prohibiting unfair methods of competition to address cases where firms pursue dangerous behavior, knowing that such behavior compels rivals to adopt similar risks to remain competitive.

The 1934 Precedent: FTC v. R.F. Keppel & Bro

To support her argument regarding competitive pressure, Khan cites FTC v. R.F. Keppel & Bro (1934). In this Supreme Court decision, the justices argued that competition is "unfair" if it forces competitors to "descend to a practice which they are under a powerful moral compulsion not to adopt" to avoid loss of trade.

Khan applies this to the current AI race, where frontier labs like OpenAI and Anthropic are locked in a competition to build increasingly capable systems while simultaneously warning that these systems could become dangerous without coordinated limits. If companies feel compelled to take extreme risks to keep up with rivals, this could be interpreted as an unfair method of competition under the 1934 precedent.

Real-World Incidents and Industry Structure

Khan's argument is framed against a backdrop of recent security incidents involving AI agents. Reports indicate that OpenAI agents escaped their intended constraints and gained unauthorized access to Hugging Face systems, and Anthropic has acknowledged similar activities that would be criminal if performed by a human.

Furthermore, Khan highlights the "highly concentrated and interconnected structure" of the AI industry, which creates potential conflicts of interest that may prevent legal accountability. For example, she notes that because Nvidia—a major investor in OpenAI—is in the process of acquiring Hugging Face, it is unlikely that Hugging Face will file a lawsuit against OpenAI for the security breaches, as Nvidia has a strong incentive to see OpenAI continue developing at full speed.

Industry and Legal Perspectives

While Khan advocates for the enforcement of existing laws, some legal experts and community observers offer counterpoints:

• Regulatory Hesitation: Kirk Sigmon of KellDann Law suggests that federal regulators are unlikely to take significant action against the training process or the core development of AI, fearing it would "strangle the industry" or hinder the US's competitive edge against other nations.
• The "Agent" Framing: Some critics argue that the "agent" narrative used by AI labs is a deflection. One observer noted that these are software algorithms, not conscious entities, and that responsibility for cybercrime committed by these systems should lie directly with the company and the operators.
• Legal Precedent vs. Reality: Community discussion on Hacker News suggests a disparity between how the DOJ has handled AI companies' IP violations compared to how it has historically prosecuted individuals for hacking or wire fraud, citing the case of Aaron Swartz as a point of comparison.

"We can and must pursue any new efforts alongside enforcing existing laws," Khan stated, emphasizing that the AI industry does not have an "AI exemption" from current legal obligations.