Leading artificial intelligence (AI) executives signed a voluntary safety accord at the White House on Tuesday, hosted by President Donald Trump. The pact sets up AI oversight without legal enforcement, even though some signatories have warned the technology could threaten humanity.
The White House Accord on Super Intelligence asks each company to monitor its own models and add outside auditors. Trump now calls AI super intelligence.
What the AI Oversight Accord Requires
The two-page document, shared on Truth Social, lists four steps. Companies are asked to monitor models internally, task a team with verifying those controls, and use outside auditors. An independent board committee would then supervise that team.
Trump called the pact morally binding when asked about enforcement. However, no credible reporting says it is not legally enforceable; the only evidence is the final paragraph, which says the steps may be codified into law or regulation over time.
In contrast, OpenAI and Anthropic reportedly came close to a legally binding mutual AI testing deal earlier this year.
Why Critics Say AI Oversight Is Thin
SpaceXAI CEO Elon Musk compared the structure to peer grading.
“grading each other’s homework, which is a lot better than if people just grade their own homework, for obvious reasons.”
Elon Musk,
Bradley Tusk, CEO of Tusk Ventures, told CNBC that almost all the same executives called for regulation two weeks ago. He argued they want a level playing field among themselves, not state power.
In a September national survey, polling firm Verasight found 63% of respondents want AI to slow down. Only 5% want it to speed up.
Meanwhile, Sen. Mark Warner, a Virginia Democrat, introduced a bill requiring rigorous AI testing and evaluation. However, CNBC reported that House Speaker Mike Johnson’s comments make congressional action much less likely.
The warnings also come from inside the industry. Anthropic CEO Dario Amodei published an essay urging a slowdown this month. OpenAI canceled the release of its Astra 6.1 model after internal tests missed safety standards.
OpenAI CEO Sam Altman said in an interview the company feels pressure to hold back unsafe models. Separately, the Federal Trade Commission (FTC) is investigating OpenAI, Anthropic, and other AI firms over consumer risks, SecurityWeek reported.
Still, voter pressure before November’s midterms may test whether lawmakers accept peer grading or write binding rules.