KOSUNI CRYPTO
Korea's Crypto Pulse, in English
$BTC $ETH $XRP $SOL $DOGE
Regulation
Neutral

Trump's AI Safety Pact Faces Scrutiny as Musk Compares It to Peer Grading

Published October 1, 2026 11:39 AM · 6 views
Trump's AI Safety Pact Faces Scrutiny as Musk Compares It to Peer Grading

Senior executives from major artificial intelligence (AI) firms have signed a voluntary safety pact hosted by President Donald Trump at the White House. The agreement aims to establish AI supervision without legal enforceability. Some signatories warned that the technology could pose a threat to humanity.

The White House Super Intelligence Accord requires companies to monitor their own models and implement external audits. President Trump now refers to AI as 'super intelligence.' A two-page document published via Truth Social outlines a four-step process: internal model oversight, verification through management teams, external audits, and supervision by an independent board.

Trump stated the accord holds moral binding power, but credible reports indicate it lacks legal enforceability. The final section notes that these measures could become legally binding over time. In contrast, OpenAI and Anthropic reportedly came close to a legally binding mutual AI testing agreement earlier this year.

Critics argue oversight is insufficient. SpaceX AI CEO Elon Musk compared the structure to peer grading. Bradley Tusk, CEO of Tusk Ventures, told CNBC that nearly the same executives had requested regulation just two weeks prior, suggesting their focus is on creating equal competitive conditions rather than responding to government power.

A September survey by VeraSite found 63% of respondents want AI speed slowed, with only 5% favoring acceleration. Meanwhile, Senator Mark Warner introduced a bill mandating strict AI testing and evaluation in Virginia, though CNBC reported that House Speaker Mike Johnson's comments have significantly reduced the likelihood of legislative action.

Warnings also come from within the industry: Anthropic CEO Dario Amodei recently published an essay urging speed reduction. OpenAI canceled the release of its Astrai 6.1 model after internal tests failed to meet safety standards. OpenAI CEO Sam Altman told interviews that the company feels pressure regarding unsafe models and has held back releases accordingly. Separately, the Federal Trade Commission (FTC) is investigating AI firms including OpenAI and Anthropic over consumer risks, according to SecurityWeek.

Korean Source

This article is an English localization of a Korean-language crypto news report. Original headline: 머스크, AI 기업 ‘서로 채점’ 주장…감시 충분한가?