Senator Adam Schiff (D-CA) outlined his views on the intersection of artificial intelligence regulation, political corruption, and the limits of executive power in a recent interview with The Verge’s Decoder podcast. Schiff, who serves on committees overseeing intellectual property, antitrust, and technology, addressed the efficacy of voluntary AI safety pledges and the potential for impeachment proceedings against President Trump.
What Happened
The interview focused on the Trump administration’s recent gathering of AI executives at the White House, where they signed a non-binding pact to evaluate their own AI systems. Schiff described the event as "jarring," noting the contrast between the executives’ public advocacy for regulation and their private compliance with the administration’s preference for voluntary measures. He argued that none of the companies wish to incur the "wrath" of the President, leading them to say what is desired in his presence while maintaining that "trust us" is insufficient for public safety.
Schiff identified a specific legal hurdle to formal regulation: the Supreme Court’s decision in Loper Bright, which overturned Chevron deference. This precedent previously required courts to defer to agency expertise when laws were vague. With that precedent removed, Schiff warned that new regulations, including the creation of an AI regulatory agency he supports, could be immediately tied up in litigation as courts now decide the interpretation of laws rather than agencies.
Why It Matters
For the AI industry and safety researchers, Schiff’s comments highlight a critical governance gap. As models advance through recursive self-improvement, Schiff noted a decreasing understanding of how AI systems make decisions and what data they access, labeling this a "national security concern of the highest order." He further cited concerns regarding the "unbridled implementation of AI at the Pentagon," suggesting a need for investigation into AI’s role in military operations to prevent "calamities."
The discussion also linked AI governance to broader political instability. Schiff cited "rampant corruption" within the administration, including the use of taxpayer dollars for presidential campaign ads and the justification of vanity projects, such as a triumphal arch, under national security pretenses. He indicated that these actions, alongside the handling of AI regulation, could influence Democratic strategies, including the potential for impeaching the President again, depending on future congressional composition following the midterms.
The Bottom Line
Senator Schiff argues that voluntary AI safety pledges are inadequate given the national security risks posed by advanced models. However, he acknowledges that the end of Chevron deference complicates the path to formal regulatory frameworks, potentially leaving new AI rules vulnerable to immediate legal challenges. The interview underscores the tension between executive preference for deregulation and legislative efforts to impose oversight on emerging technologies.