Bessent told the House that AI labs should stay liable for what they build

What the labs were asking for sat outside either official's power to grant or deny. A liability exemption for model developers takes a statute, and Bessent runs Treasury ^[1][4]. Ferguson chairs the FTC, but he said he was giving a personal opinion, and he noted that President Trump sets federal AI policy ^[10]. Two senior figures declined to carry the industry's case.

The default stays in place: whatever existing law already assigns. "The best way to guarantee safety is that the creators are liable for what they build and generate," Bessent said ^[3]. He was answering Rep. Juan Vargas, D-Calif., at a House Financial Services Committee hearing devoted to IMF oversight ^[4]. He declined to identify a draft regulation or bill that he believed would improve AI safety ^[6]. The House was due to leave Washington on Thursday with AI bills stalled ^[18].

OpenAI removed the factual premise of the waiver request. Chris Lehane, its chief global affairs officer, said the company had worked with Anthropic and Google DeepMind on safety for several weeks without needing an antitrust waiver ^[13]. "There are many instances over time where companies are trying to help each other on safety," Lehane said, citing the airline industry ^[14]. The same day, OpenAI backed the independent-verification provision of the FRONTIER Act, its first endorsement of a specific federal mandate requiring outside safety assessors ^[12]. Implicator.ai reported that existing antitrust doctrines already allow coordination to prevent catastrophic risks, including information sharing on cybersecurity ^[15].

Ferguson's stated objection was that the paired requests could create barriers protecting established companies from challengers ^[9]. Verification obliges a lab to open its own models to an outside assessor; the waiver the industry sought would have let rival labs coordinate a slowdown with each other ^[11]. Cohere co-founder Aidan Gomez made the same point on Sunday: "If this is the most consequential technology in human history, then the rules for it cannot be written by a small group of commercially aligned companies behind an antitrust waiver," he wrote ^[16].

The industry is asking for different things. Nvidia chief executive Jensen Huang opposed new laws on Tuesday. "Run as fast as you can," he said. "But if you feel at any given point in time the company's out of control or the product's not going to be safe, take a pause and make sure you get it right" ^[17].

One number keeps circulating in the argument. An Elon University survey of 1,505 U.S. adults found 69% expect AI to weaken human bonds, about 1,038 respondents ^[19][20]. It was fielded in May, before this month's slowdown debate ^[19].

Bessent paired his answer with a second position. He called for more open-source models built in the United States, and said: "We can't let these large labs have regulatory capture because that will stop innovation" ^[5].