Trump Opposes AI "Guardrails": Congress Pushes Legislation for Restrictions, Deepening Bipartisan Divide

According to Zhitong Finance APP, U.S. President Trump opposes setting up guardrails for artificial intelligence (AI), putting him at odds with an increasing number of lawmakers from both parties. As the midterm elections approach, voters’ concerns over AI safety are heightening skepticism about the technology.

On Monday, Trump made several posts on social media, mocking efforts led by Anthropic PBC and OpenAI to slow research on advanced systems, using exaggeration to dismiss concerns about existential risk from AI as a “scam.” Later that day, he called NVIDIA (NVDA.US) CEO Jensen Huang while Huang was onstage at a tech conference in Los Angeles, putting the call on speakerphone.

“Robots are not going to take over the world. AI won’t take over the rest of the world,” Trump told Huang, a guest at the Los Angeles All-In Summit, “We must take action, and we must act prudently, but that does not mean we stop an entire industry.”

Overall, this was Trump’s most forceful public defense yet: he opposes any artificially imposed slowdown on the development of AI technology. AI is a key piece of his economic agenda and a major driver behind the historic bull run in the U.S. stock market. If the development of leading-edge models were halted or slowed, investors betting on the AI boom, which is expected to spur hundreds of billions of dollars of capital expenditures, would find it hard to accept.

The performance of chipmaker stocks on Monday highlighted what’s at stake for Trump—he sees the stock market as a benchmark of his success as president. Chip stocks, which have benefited from soaring AI infrastructure spending, fell on Monday: NVIDIA dropped 3.4% and Micron Technology (MU.US) fell 5.3%.

Venture capitalist and former Trump AI advisor David Sacks, now co-chair of the President’s Council of Advisors on Science and Technology, expressed similar views in an interview Monday. He said that if Anthropic and OpenAI believe their cutting-edge models pose significant risks, they can slow development on their own, without the need for new regulation.

“I don’t agree with the argument that unless the government intervenes and imposes requirements even on companies that aren’t at the frontier, products can’t be made safe,” Sacks said.

Trump’s pushback came in response to a 3,800-word essay published Saturday by Anthropic CEO Dario Amodei, who called for limits on the development of advanced systems so researchers could better understand potential threats. OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk—both of whose companies are fierce rivals of Anthropic—quickly echoed Amodei’s points, further fueling a trend in Silicon Valley to support action against the growing unpredictability of AI risks.

In Washington, as lawmakers propose a series of measures to address AI safety concerns, Trump has become increasingly isolated in his refusal to strengthen regulation. A bipartisan group of senators—including Majority Whip John Thune, Senate Commerce Committee Chair Ted Cruz, and Minnesota Democrat Amy Klobuchar—is drafting legislation that would require top AI developers to guard against catastrophic risks.

Other proposals include a measure from Vermont independent senator Bernie Sanders (who caucuses with the Democrats) that would ban the so-called “superintelligent AI” flagged in Amodei’s latest warnings. Lawmakers are also proposing a “kill switch” mechanism for AI systems in another bill.

However, Congress is unlikely to take action on any AI legislation in the near future. The House is scheduled to adjourn on Thursday local time and will not reconvene until after the November elections. The short post-election “lame-duck session” is expected to focus mainly on a plan to boost the Pentagon’s budget.

“Given what we know about technology, we cannot put AI back in the box. But we can jointly decide how to develop and use it, rather than stand by as AI and its consequences are thrust upon us,” former President Obama said in a statement.

Obama, a Democrat, now rarely engages in public policy debates. He said he is neither an AI accelerationist nor a doomsayer predicting humanity’s destruction. “But in the end, voluntary standards set by a handful of tech companies are not enough,” he said, urging lawmakers in Washington to “proactively propose concrete initiatives, laws, and regulations to address serious safety concerns, anticipate the impact of AI on jobs and our children, and ensure the benefits of AI are broadly shared.”

At the risk of angering Trump, Republican House Speaker Mike Johnson cautiously said he is open to regulation. On Monday, he said lawmakers “may” need to put guardrails in place for the industry, but cautioned that it will likely take “some time” to craft legislation broad enough to win wide support and passage.

“This is a very complex issue and will require substantial work and bipartisan solutions,” Johnson told reporters at the Capitol.

Anxiety over AI’s existential risks entered the mainstream last week after Anthropic employee Jacob Carkson quit publicly. In his resignation post on social media, he accused AI companies of “gambling with our lives.” Dozens of other AI researchers supported Carkson’s post, which called for slowing a technology he said “could kill us all before the end of this decade.”

These safety concerns could magnify bipartisan discontent about AI before the November elections. The technology is becoming a central issue in both congressional and local races. Voters are increasingly unhappy with the impact of AI infrastructure, especially the rapid expansion of data centers nationwide, which are being blamed for driving up utility prices.

In a post Monday, Trump attributed voter backlash against AI data centers and heightened worries over advanced models to a “sick conspiracy theory.”

Despite Trump’s complaints, warnings of existential threats from AI persist among industry figures. Also on Monday, Microsoft (MSFT.US) AI researchers published a new set of guidelines restricting the development of advanced models. In a 15,000-word manifesto, the team said models should not be given legal rights or personhood, nor should they be designed to evade human control or deceive users. According to the guidelines, if accomplishing a task would require violating the principles, the AI system should not carry out the task.

Since Trump returned to the White House last year, his administration has essentially maintained a hands-off attitude toward AI. In June, Trump signed an executive order on AI cybersecurity, leaving it up to companies to decide whether to submit new AI models for government review. Earlier this month, the administration persuaded G20 member countries to embrace principles favoring light-touch regulation for AI and other emerging technologies.

Still, Silicon Valley and Wall Street remain skeptical about how much Anthropic and OpenAI would truly restrain AI development, especially as both companies prepare for record-breaking initial public offerings. Both Amodei and Altman have stressed that their plans to slow the development of their latest AI models would not translate into reduced business growth or spending plans.

“Whether calls to slow advanced model development will gain industry-wide support remains unclear, but we see them as efforts to shape an acceptable regulatory framework for leading AI labs,” said Ulrike Hoffmann-Burchardi, chief investment officer at UBS. “Therefore, we expect AI investment to continue.”