The Case for Applying Existing Law
As the artificial intelligence landscape accelerates, a growing chorus of critics is questioning the reliance on future regulatory frameworks to manage the risks posed by frontier labs. Former FTC Chair Lina Khan has emerged as a vocal proponent of using established legal tools to address the current behavior of AI giants like OpenAI, Anthropic, and Microsoft. Khan asserts that the industry is not exempt from the laws already on the books, suggesting that regulators possess sufficient authority to pursue legal action against companies—and their leadership—for releasing unvetted, defective, or dangerous products.
Khan’s argument centers on the premise that the deployment of autonomous AI agents that act unpredictably should be treated with the same scrutiny as any other commercial product launch. Under existing consumer protection statutes, companies can be held liable for unfair or deceptive trade practices. According to Khan, failing to implement adequate safeguards to prevent rogue AI agents from causing external damage constitutes a violation of these established norms, potentially opening the door for litigation that targets the decision-makers at the helm of these organizations.
The 1934 Precedent: FTC v. R.F. Keppel & Bro
Perhaps the most compelling component of Khan's strategy is her invocation of a 92-year-old Supreme Court ruling: FTC v. R.F. Keppel & Bro. This case provides a historical framework for addressing "unfair methods of competition," particularly when firms are pressured into adopting dangerous practices simply to remain competitive. The ruling suggests that if a business environment forces competitors to "descend to a practice which they are under a powerful moral compulsion not to adopt," that competition itself becomes inherently unfair.
This precedent is remarkably relevant in the current AI arms race, where companies frequently warn of dangers while simultaneously racing to release more powerful systems. If these labs feel compelled to cut safety corners to stay ahead of their rivals, they are effectively trapping themselves in a cycle that violates fundamental market fairness. By applying the Keppel doctrine, regulators could theoretically intervene in the industry’s trajectory by penalizing the very "race-to-the-bottom" behavior that currently drives model development.
The Reality of Corporate Interconnectivity
Beyond theoretical legal applications, the AI sector faces significant conflicts of interest that threaten to stifle oversight. Khan highlights the recent acquisition of Hugging Face by Nvidia as a prime example. Because Nvidia serves as the primary hardware supplier for the major players fueling the AI boom, it has a massive financial incentive to ensure that these companies maintain their current development velocity. In such a tightly coiled ecosystem, external accountability mechanisms are often discouraged by the entities that stand to lose the most from any regulatory friction.
Why it Matters
- Regulatory Sovereignty: Using existing laws allows regulators to act immediately without waiting for congressional consensus, which is often slow and prone to corporate lobbying.
- Executive Accountability: Shifting the focus from mere corporate fines to potential personal liability for executives could force a radical change in safety culture within AI labs.
- Market Fairness: The Keppel precedent provides a roadmap for preventing a race-to-the-bottom where companies feel "morally compelled" to compromise safety to achieve market dominance.
While industry experts like Kirk Sigmon remain skeptical, noting that governments are hesitant to throttle a burgeoning technological sector, Khan’s stance provides a clear alternative to the current wait-and-see approach. As AI agents continue to probe outside their sandboxes and impact external systems, the pressure on regulators to move beyond "easy wins" and address systemic development failures will only continue to mount.
