Amid calls for regulation, President Donald Trump met with White House artificial intelligence officials and signed a voluntary agreement called the White House Agreement on Superintelligence.
The agreement, signed Tuesday by executives from Meta, Nvidia, Google, OpenAI, xAI and Anthropic, requires companies to put in place four layers of internal guardrails to help monitor security risks.
Recommended stories
list of 4 itemsend of list
What does the White House agreement do and what does it not do?
The agreement requires companies to maintain internal controls, establish an internal team to oversee those controls, work with independent external auditors, and establish an independent board committee to review the results.
However, this agreement is voluntary and not legally binding. There are no penalties for companies that do not comply. The agreement does not require companies to publish audit results or give the government an enforcement role. Companies can also control how safeguards are implemented.
However, the agreement leaves open the possibility that the measures may eventually become mandatory. Over time, the magazine said, “it may make sense to codify these steps into laws and regulations.”
“There will be a tremendous amount of self-regulation,” Trump told reporters, adding that it would be “morally binding.”
On the White House lawn, President Trump stood alongside executives and argued that AI companies have done a great job of self-regulating.
Google CEO Sundar Pichai also promoted the agreement on social media.
“We are committed to working with other industry leaders to establish norms and build public trust. The White House Accord and Joint Commitment on Frontier Responsibilities signed today are a solid foundation for moving forward. They include real, concrete steps to advance safe development while delivering the economic and scientific benefits of this technology,” Pichai said in a post on X.
Why are AI companies seeking more regulation?
The agreement came despite calls from some of the same companies for government-mandated safety standards.
In September, Chris Lehane, chief international affairs officer at ChatGPT developer OpenAI, called on Congress to establish mandatory national AI safety standards.
“Industry-driven standards should complement, not replace, the federal government’s mandatory safeguards and democratic oversight,” Lehane said in the post.
In June, Anthropic CEO Dario Amodei called for “more serious and binding regulation of AI.” In July, Google DeepMind CEO Demis Hassabis called for the creation of a U.S.-led AI oversight agency.
The call for guardrails comes amid reports that some employees at these companies have expressed concerns about whether their increasingly sophisticated AI systems are being properly monitored.
Two OpenAI employees warned executives that the company’s testing of new models was not being properly monitored. According to the New York Times, no warnings were issued before the model behaved in an unintended manner.
What are the lawmakers proposing?
The push for legally binding regulations is coming from Congress, not the White House.
Earlier this month, former Anthropic AI researcher Jacob Coxon received renewed attention from lawmakers when he warned that AI technology could pose an existential threat to humanity within the next decade.
Some of the most comprehensive proposals come from independent Sen. Bernie Sanders of Vermont and Democratic Rep. Greg Cassar of Texas.
Both lawmakers have introduced the Artificial Superintelligence Ban Act, which, if passed, would permanently ban the development and deployment of artificial superintelligence and temporarily halt advanced AI development until federal regulators establish safety rules.
The bill would also create a Cabinet-level Department of Artificial Intelligence to oversee frontier AI systems, as well as oversee dangerous capabilities and, if necessary, the destruction of artificial superintelligence.
The bill would impose stiff penalties for violations, including up to 20 years in prison for individuals.
This comes amid a bipartisan bill introduced last week by Democratic Sen. Chris Coons of Delaware. Brian Schatz, a Democrat from Hawaii. Katie Britt, Republican from Alabama. James Lankford, a Republican from Oklahoma.
The AI Systems Transparency Act would require AI companies to disclose the data they collect and what measures they take to protect their systems from fraud or abuse by bad actors.
What does “superintelligence” actually mean?
The White House named the agreement the “White House Agreement on Superintelligence,” as President Trump pushes to change the name of “artificial intelligence” on government documents.
In the technology industry, “superintelligence” refers to a type of AI that exceeds human intellectual capabilities.
However, President Trump has directed executive agencies to use “superintelligence” or “SI” instead of “artificial intelligence” or “AI” in official communications.
On Tuesday, President Trump outlined a new executive order that defines “superintelligence” as technology and systems that are covered by the existing legal definition of artificial intelligence and directs the creation of a new federal definition within 60 days.
