Scott Bessent Slams AI Alarmism, Urges Industry to Own Risks as US Plans China AI Dialogue

Scott Bessent Rejects AI Alarmism Without Solutions

US Treasury Secretary Scott Bessent has sharply criticized AI industry leaders who warn of catastrophic risks but fail to propose actionable solutions, calling their approach a failure of leadership1

1

. In an interview with Axios published Saturday, Bessent stated that AI alarmism without solutions by some of the AI industry community represents poor leadership2

2

. The Treasury Secretary went further, comparing AI executives seeking federal guardrails to Hannibal Lecter, the fictional serial killer, saying it's like the character pleading "Stop me before I kill again"2

2

. His comments directly address recent calls from prominent AI industry leaders including Anthropic CEO Dario Amodei, OpenAI chief Sam Altman, and Elon Musk, who in September urged the industry to slow development of its most advanced models due to growing existential risks for humanity3

3

. Microsoft co-founder Bill Gates also warned the technology is powerful enough to cause "a billion deaths" and called for government regulation3

3

.

AI Industry Must Own Its Risks Through Self-Regulation

Bessent made clear that the responsibility for managing AI risks and opportunities belongs squarely with the companies developing the technology. "It is the people at the labs who have to accept responsibility, and I agree with that," he told Axios1

1

. When asked about AI executives seeking federal regulation, Bessent replied bluntly: "Well, then they should slow down"1

1

. His position aligns with President Donald Trump, who has rejected calls for new federal regulation and labeled increasingly dire warnings about risk to humanity a "hoax"3

3

. During a recent White House meeting with top industry executives including Elon Musk, Microsoft CEO Satya Nadella, Meta's Mark Zuckerberg, Palantir CEO Alex Karp, and OpenAI President Greg Brockman, Trump emphasized self-regulation over government intervention3

3

. Trump told reporters that "there's a belief that there should be tremendous self-regulation" and endorsed using outside auditors to assess AI systems' safety through a non-binding accord with Silicon Valley leaders3

3

. The administration produced a voluntary safety pact with AI CEOs, meaning the commitments are not binding law4

4

.

Balancing Innovation With Safety in AI Governance

Bessent's stance reflects the Trump administration's approach to AI governance: accelerate development while maintaining US leadership over China, but do so through industry self-policing rather than federal regulation. "We want safe acceleration," Bessent said3

3

. The Treasury Secretary noted that "the labs have been very cooperative," citing OpenAI's recent decision to shelve a new model after internal tests raised questions about whether it would follow user instructions4

4

. However, not everyone in the AI industry is convinced this approach adequately addresses safety concerns. Sarah Heck, Anthropic's head of public policy, stated on X that "You can't do safety from second place"4

4

. David Robinson publicly announced his resignation from OpenAI and criticized the company's rapid development culture, arguing that OpenAI is not exercising enough caution as its AI systems become more capable4

4

. House Speaker Mike Johnson told CNN's Jake Tapper that AI executives "need to take personal responsibility first, and we need to get them all in a room at the same time and agree on what the guardrails are"2

2

.

US-China Discussions on AI Risks and Warning System

Despite the competitive rhetoric, Bessent revealed that the US plans to propose a bilateral notification system with China aimed at reducing uncertainty around AI incidents. The system would enable both governments to share warnings about serious incidents and respond before an AI problem escalates or spreads globally1

1

. Bessent and Chinese Vice Premier He Lifeng discussed artificial intelligence during talks in New York in September, agreeing to create a "US-China AI dialogue" to develop a common understanding of the technology's benefits and threats3

3

. The discussions covered risks such as rogue agents and non-state actors using AI for cybersecurity or biological threats3

3

. Bessent warned that Chinese models can be "80%, 90% as powerful" as US ones but come without guardrails and can be imported anywhere in the world1

1

. Chinese President Xi Jinping told Donald Trump during his Washington visit that China and the US can keep discussing AI, exchange views on its risks and benefits, and work together to prevent its misuse3

3

. However, this openness has limits. Trump stated last week he does not want an AI safety deal with China1

1

. The tension between maintaining US competitiveness and managing global AI risks will shape how the administration approaches AI governance in the coming months, with companies expected to demonstrate they can police themselves effectively or face potential pressure for stronger oversight.