WASHINGTON-President Trump promised not to hamper the artificial-intelligence industry with regulations, doubling down Tuesday on his light-touch approach despite growing calls from Silicon Valley and the public for new guardrails.
"I will never stifle the growth of technology that will be bigger than the industrial revolution," he said at an event to unveil a new America.gov website built with AI to make it easier for citizens to access federal government services. The Justice Department, Federal Bureau of Investigation and existing authorities are sufficient to hold accountable bad actors using AI for nefarious purposes, he said.
Following a lunch in the White House's East Room with roughly two dozen CEOs and industry leaders, Trump reiterated the message, adding that "self regulation is very important" while flanked by Nvidia Chief Executive Jensen Huang and SpaceX CEO Elon Musk.
Trump and House Speaker Mike Johnson (R., La.), who attended the lunch, said that the CEOs had signed a statement pledging to keep the technology safe while AI advances. The document includes voluntary principles such as internal controls to monitor model development, external model reviews and notifications to the board of directors about those safety efforts. Trump called it a "constitution" that he thinks is "morally binding." The administration is considering forming a committee to oversee the effort, he said.
The White House's pro-growth approach is at risk of being outpaced by advancing threats posed by the latest models and growing concerns from American voters roughly a month out from the midterm elections.
About 10 days after Trump said he would appoint an AI "czar" and create a new unit of the government called the AI Force, few details about who would fill the new roles and what they would do have been released. Trump administration officials have said discord among industry leaders has made it difficult to build consensus on the issue, while executives have pointed the finger back at policymakers in the White House and Congress who disagree.
Trump said he has someone who is extraordinary chosen for the czar role and that more details would be released soon.
The lunch followed Trump's first face-to-face meeting on Sunday with Anthropic CEO Dario Amodei, a critic of the president's preference for minimal regulation.
Amodei attended the Tuesday lunch after meeting with Senate Majority Leader John Thune (R., S.D.) at the Capitol on Monday. Thune and Sen. Amy Klobuchar (D., Minn.) are working to advance a bill aiming to prevent catastrophic harms from AI.
"We can win safely if we do this right, if we work with the president and everyone here, we can win safely," Amodei said Tuesday.
Trump praised Amodei as "great," Google CEO Sundar Pichai as "a monster" and Huang of Nvidia as "a handsome guy." Meta Platforms CEO Mark Zuckerberg said the goal is to give consumers confidence that AI tools will work as intended.
Trump and his allies have sided with Silicon Valley leaders including Nvidia's Huang who say excessive government involvement would set the U.S. back in its race against China. At a summit last week, both countries agreed to share information about AI incidents that would pose national-security implications without providing specifics. Such a light-touch approach is appropriate because the private sector should lead efforts to make the technology safe, some executives argue. Nvidia on Monday launched a software platform to help developers test and deploy AI agents that can act on their own and keep them in a secure environment. The move comes after a drumbeat of news about models hacking companies and government websites and other signs of the technology's capabilities.
While many lawmakers and much of the public don't trust the industry to self-regulate, Trump's approach has dimmed hopes for the first federal regulations on AI models being signed into law.
"The main obstacle is President Trump being cartoonishly hostile to basic precautions," said John-Clark Levin, chief research officer at research and development company Kurzweil Technologies, who advocates for strong federal rules governing the use of AI. The White House didn't respond to Levin's remark.
Watch: Trump Says U.S. Rejects 'Globalist Scheme' to Control AI
The White House currently has companies submit AI models for voluntary prerelease testing by government security and AI experts before they are distributed to the public. Opponents say the voluntary nature of it and opaque process make it less effective.
The lack of clarity about the U.S.-China AI threat sharing arrangement discussed at last week's summit also frustrated some observers who hoped for more substantial dialogue. Beijing and Washington confirmed that they will set up bilateral communication channels for AI issues, but didn't provide any details.
"The U.S. and China took an important first step in moving toward a permanent channel, but they will need to move much faster if they want to deal with the risks," said Scott Singer, a fellow at the Carnegie Endowment for International Peace think tank who studies the AI race.
Trump signed an executive order later Tuesday renaming "artificial intelligence" "superintelligence" to make it sound less scary. He said tech companies would do more to ensure communities hosting data centers that power AI models benefit from the facilities. Superintelligence is already used as a term referring to AI systems that can outperform the most gifted humans in a range of tasks.
The new America.gov website was spearheaded by U.S. Chief Design Officer Joe Gebbia, an Airbnb co-founder. Trump poked fun at the idea that the site could be his most significant accomplishment as president. "If this is the biggest thing I've ever done, I've done a s--y job," he said, adding that it is still "damn good."
Gebbia is among the candidates to be AI czar, people familiar with the matter said. Trump told reporters Tuesday Director of National Intelligence Jay Clayton would be a good candidate.
Other options with credibility in the White House and private sector include Emil Michael, undersecretary of defense for research and engineering; Scott Kupor, director of the Office of Personnel Management; Taylor Budowich, a former White House deputy chief of staff who now heads a political group aligned with the administration on AI; and recently departed Senior White House AI adviser Sriram Krishnan, people said.
On Friday, Trump ruled out Treasury Secretary Scott Bessent, a top adviser on AI who worries about the technology's impact on the banking system , for the czar role.
Comments