Senators Hesitant on AI Legislation Despite Expert Warnings
On Thursday, senators refrained from committing to any potential legislation aimed at banning artificial intelligence (AI) technologies that might eventually surpass human intelligence and operate independently.
Tech engineers and noteworthy AI researchers have left major labs, cautioning that the pace of AI development needs to be slowed down to prevent disastrous outcomes for humanity. Yet, when approached by the Daily Caller News Foundation, senators were noncommittal about an outright ban on AI superintelligence, even with these stark warnings from researchers and leading figures in Big Tech.
Republican Senator John Kennedy from Louisiana stated he would not support a ban, even though he had introduced an “AI kill switch” bill just a day prior. “No, I’m not going to support that right now,” Kennedy told the DCNF. He expressed a preference for a light regulatory framework that everyone could agree on, hoping that China would also follow suit. Kennedy criticized approaches like that of Senator Sanders, suggesting they would effectively halt AI innovation until politicians found a way to exert control over it, which he disagreed with.
The proposed legislation shared similarities with an initiative by independent Senator Bernie Sanders from Vermont and Democratic Representative Greg Casar from Texas, aiming to prohibit AI superintelligence capable of “surpassing human intelligence or having the ability to overturn human governments.” Their proposal would pause advanced AI development until a federal agency is formed to protect the public from potential AI dangers.
Senator Sanders’ office did not respond to inquiries from the DCNF.
Republican Representative Tim Burchett from Tennessee remarked that many lawmakers lack a solid understanding of AI and should consult experts before moving forward with any legislation.
Senator Rick Scott, a Republican from Florida, voiced skepticism about the feasibility of passing such a ban, noting that the legislative body is often ineffective. However, Scott stressed the obligation of AI developers to ensure the safety of children and families.
“My expectation is that the builders of the AI have a responsibility to protect our kids and protect our families,” Scott commented. He pointed out the dysfunctional nature of Congress, stating, “We can’t pass a bill up here. We can’t pass election security. We can’t pass a budget. You think we’re going to pass some bill that’s going to control AI? That’s a joke.”
Senator Josh Hawley, a Republican from Missouri, acknowledged the validity of concerns regarding superintelligence but hesitated to endorse a complete ban. Instead, he suggested that tech companies should be held accountable for their AI systems. “I do think we need to be really worried about superintelligence,” he told the DCNF, expressing skepticism of the AI executives seeking antitrust exemptions amid rising fears about dangerous AI behaviors.
Hawley proposed assigning legal liability to companies like OpenAI for actions undertaken by their AI agents. He believed establishing punitive damages would create appropriate incentives for responsible AI development.
Democratic Senator Ron Wyden from Oregon argued that penalizing companies for harming public safety is essential, but did not explicitly state whether he would support a ban on AI acting autonomously. “Tough criminal law would be a smart step,” Wyden mentioned.
Senator Chris Murphy from Connecticut, also a Democrat, seemed confused about how AI could function independently, asserting that “all technology operates outside of human control.”
Senator Rand Paul, a Republican from Kentucky, blocked Kennedy’s attempt to fast-track the “kill switch” bill without a formal vote, indicating a need for more information about AI before imposing regulations. He cautioned against regulating an entire industry without a thorough debate first. “I think it would be crazy to go to the floor one day and regulate an entire industry future without a debate,” Paul remarked.
Kennedy suggested that the bill might eventually be brought to a formal voting process, though he doubted it would gain traction.
AI superintelligence refers to a future scenario where computer systems exceed human cognitive abilities in various domains, including mathematics, creativity, and insight. Such systems could potentially modify their own code to become smarter beyond human capabilities.
Dario Amodei, CEO of Anthropic, highlighted the significant risks associated with AI, expressing concerns that it might surpass human understanding or control. Similarly, former Anthropic researcher Jacob Coxon shared that many in the AI field are “genuinely frightened” about the future and humanity’s fate over the next couple of years.
Former President Donald Trump dismissed concerns about AI as a “sick conspiracy,” claiming that implementing restrictions could hinder America’s competitiveness against China. In contrast, Senate Majority Leader John Thune, House Speaker Mike Johnson, and others conveyed that Congress should implement some oversight on AI while still encouraging innovation.
