“Tremendous self-policing”: Trump unveils AI safety agreement
‘Tremendous self-policing’: President Trump unveils voluntary AI safety accord with tech leaders
WASHINGTON —
President Donald Trump and leading technology executives announced a new voluntary industry agreement focused on artificial intelligence (AI) safety. It emphasized self-regulation by companies rather than new government rules.
After the closed-door meeting, Trump emerged with prominent industry leaders standing behind him to unveil the White House accord aimed at ensuring the responsible development of AI.
“There is going to be a tremendous self-policing aspect,” Trump said.
The agreement included measures such as stronger internal controls, the use of outside auditors and evaluators, and independent reviews of safety reports by company boards.
Mark Zuckerberg, the CEO of Meta, described the accord as a starting point for the industry.
“The idea isn’t that this is the only thing that we will ever do. It is that this is a start and an accord that the whole industry could come to,” Zuckerberg said.
Earlier in the day, Trump rejected calls for new government guardrails on AI, and said he would “never stifle” development of the technology.
Dario Amodei, the CEO of Anthropic, has been a vocal advocate for addressing AI risks, said the broader question of how to manage those risks remained unresolved.
“But I think the technology has very real risks. And the mechanism, how we address those risks, is still under discussion. But we all need to work. We all need to work together to make sure that we can win and we can win safely,” Amodei said.
Amodei has previously warned that advanced AI could pose a risk to human existence. OpenAI recently delayed the release of its latest model following internal safety testing.
Trump also highlighted the importance of maintaining the United States’ competitive edge over China in AI development.
“They’re quite a ways back, and we have to keep it that way,” Trump said.
When asked about the binding nature of the voluntary agreement, Trump described it as “morally binding.”
Separately, the government launched a new AI-powered website called “America.gov,” which currently functions as a search assistant for government information. The site pulls data from federal agencies using only government sources.
The administration plans to expand its capabilities next year to handle tasks such as changing names, checking benefits, and completing other government services without requiring users to navigate multiple agency websites.
