Trump opposes AI “protective fence”: Congress pushes legislation to set limits, differences between the two parties deepen

The Zhitong Finance App learned that US President Trump opposes setting up a protective fence for artificial intelligence (AI), which has divided him with more and more lawmakers from both parties. As the midterm elections approach, voters' concerns about AI security have intensified their doubts about the technology.

On Monday, Trump posted multiple posts on social media, scoffing at Anthropic PBC and OpenAI's efforts to slow down the development of cutting-edge systems, using exaggerated rhetoric to dismiss concerns about AI's existential risks as a “scam.” Later that day, he impromptu telephoned Nvidia (NVDA.US) CEO Wong In-hoon, who was on stage at a technology conference and set the call hands-free.

“Robots won't take over the world. “AI won't take over the rest of the world,” Trump said to Hwang In-hoon, who is a guest at the All-In summit in Los Angeles. “We have to act and act carefully, but that doesn't mean we need to shut down an industry.”

Overall, this is Trump's strongest public defense to date: he opposes artificially slowing the pace of AI technology development. AI is at the core of its economic agenda and an important driver behind the historic bull market in the US stock market. Once cutting-edge big model development is suspended or slowed, investors who bet that the AI boom will drive hundreds of billions of dollars in capital expenses will inevitably find it difficult to accept.

The chipmaker's stock price performance on Monday highlighted the stakes Trump faces — he sees the stock market as a benchmark for success in his presidency. Chip stocks, which benefited from a surge in AI infrastructure spending, fell on Monday, with Nvidia falling 3.4% and Micron Technology (MU.US) falling 5.3%.

Venture capitalist David Sachs, former head of Trump's AI affairs, and now co-chairman of the President's Science and Technology Advisory Council, expressed similar views in an interview on Monday. He said that if Anthropic and OpenAI consider their most advanced models to pose a serious risk, they can completely slow down development on their own without the need for new regulations to push forward.

“I don't agree with this statement — that unless the government steps in and imposes requirements on companies that aren't even at the cutting edge, they can't guarantee product safety,” Sachs said.

Trump's counterattack stems from Anthropic CEO Dario Amodii's 3,800-word paper last Saturday urging restrictions on the development of cutting-edge systems so researchers can better understand potential threats. OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk — the two companies are fierce rivals with Anthropic — quickly agreed to Armodei's views, further fueling the trend within Silicon Valley to support measures to address the increasingly unpredictable risks of AI.

In Washington, Trump became increasingly isolated in refusing to strengthen regulations as lawmakers proposed a series of measures aimed at addressing AI security concerns. A cross-party group of senators — including Majority Leader John Tune, Senate Commerce Committee Chairman Ted Cruz, and Minnesota Democratic Senator Amy Klobucher — is drafting legislation requiring top AI developers to guard against catastrophic risks.

Other proposals include a measure proposed by Vermont Independent Senator Bernie Sanders (allied with the Democratic Party) calling for a ban on the so-called “super-smart AI” that sparked Amoday's latest warning. Lawmakers also proposed the idea of setting up an “emergency kill switch” for AI in another bill.

However, Congress is unlikely to act on any AI legislation anytime soon. The House of Representatives is scheduled to adjourn on Thursday local time, and will not resume its session until the November general election is over. The short “lame duck session” after the election is expected to focus on a plan to boost the Pentagon's budget.

“Given what we know about technology, we can't put AI back in the box. But we can decide together how to develop it and how to use it, rather than stand idly by and let AI and its consequences befall us,” former US President Barack Obama said in a statement.

As a Democrat, Obama is now rarely involved in public policy debates. He stated that he is neither an AI accelerator, nor an apocalyptic who predicts the destruction of humanity. “But at the end of the day, voluntary standards set by a few tech companies aren't enough,” he said, and urged lawmakers in Washington to “come forward with specific proposals, laws, and regulations to address serious safety concerns, predict the impact of AI on employment and our children, and ensure that the benefits of AI are widely shared.”

At the risk of angering Trump, Republican Speaker of the House Mike Johnson cautiously expressed openness to regulation. He said on Monday that lawmakers “may” need to put a protective fence on the industry, but cautioned that it will probably take “some time” to enact legislation sufficient to gain broad support and pass.

“This is a very complicated issue that requires a lot of work and a joint solution from both parties,” Johnson told reporters at the Capitol.

Anxiety about the risks of AI flooded into the mainstream last week due to the high-profile departure of Anthropic employee Jacob Coxon. In his resignation post shared on social media, he accused the AI company of “gambling our lives.” Dozens of other AI researchers supported Coxon's post, which called for slowing down a technology he said “could kill us all before the end of this decade.”

These security concerns are likely to fuel bipartisan dissatisfaction with AI before the November elections. This technology is becoming a central topic in congressional and local elections. Voters are increasingly dissatisfied with the impact of AI construction, particularly the rapid expansion of data centers across the US, which is blamed for driving up utility prices.

In a Monday post, Trump blamed voters' backlash against AI data centers and heightened concerns about cutting-edge models on a “sick conspiracy theory.”

Despite Trump's complaints, industry people continue to warn that AI is a sexual threat. Also on Monday, AI researchers at Microsoft (MSFT.US) released a new set of guidelines to limit the development of cutting-edge models. In a 15,000-word declaration, the team said the model should not have legal rights or personality, and should not be designed to evade human control or deceive users. According to the declaration, if completing a task requires violating the guiding principles of AI, the system should not perform that task.

Since Trump returned to the White House last year, his administration has largely taken a non-intrusive attitude towards AI. In June, Trump signed an AI cybersecurity executive order, leaving companies to decide for themselves whether to let the government review their new AI models. Earlier this month, the government persuaded G20 members to accept principles advocating light-touch regulation of AI and other emerging technologies.

However, Silicon Valley and Wall Street are still skeptical about the extent to which Anthropic and OpenAI will actually restrain AI development, especially as the two companies are preparing a record initial public offering. Both Amodei and Altman have made every effort to stress that their plans to slow down the development of the latest AI models will not translate into business growth or a contraction in spending plans.

“The appeal to slow advanced model development is still uncertain, but we think they are more aimed at shaping an acceptable regulatory framework for leading AI laboratories,” said Ulric Hoffman-Burchardi of UBS's Chief Investment Office. “Therefore, we expect AI investments to continue.”