Elon Musk called on Monday for the leading artificial intelligence companies to run safety tests on each other’s models before public release, speaking at the All-In Summit.
The SpaceX chief executive said his xAI business, OpenAI, Anthropic, Google, Meta and “three or four of the leading Chinese companies” should let rivals run a “test harness” on their systems to evaluate safety.
“Instead of grading your own homework, you would at least have competitors grading your homework and raising the alarm if they see concerns,” Musk said. He acknowledged the method was imperfect but added, “The odds that you will find issues are dramatically greater.”
The debate around regulating AI has intensified. Tech leaders backed Anthropic chief executive Dario Amodei‘s proposal for a slowdown, a rare show of unity. The push came amid warnings about the technology last week, sparked by a post from Jacob Coxon, who announced that he quit his job at Anthropic.
Coxon, who also worked at OpenAI, wrote that the leading labs are “gambling with our lives.” Evan Hubinger, an alignment lead at Anthropic, backed the statement, writing that he believed AI could “kill all humans.” He wrote that he personally thinks the probability is above 10% within the next decade.
The Trump administration has pushed back on calls for greater limits. President Donald Trump posted on Truth Social on Monday referring to fears over AI as a “hoax” and a “scam.” National Economic Council Director Kevin Hassett told CNBC on Tuesday that the private sector was the “right place” to address concerns, adding the government would use “law enforcement when necessary.”
Trump has argued that pacing AI development would hand China an advantage, a concern Amodei echoed, calling it the “toughest dilemma.” A spokesperson for China’s Foreign Ministry on Monday called the push for a slowdown “fear mongering,” according to a Reuters translation.
Musk said rival labs competing with xAI, which SpaceX acquired in February, had not agreed to his proposal. Amodei’s essay suggested labs could instead accept third-party assessments from “embedded evaluators” who would verify safety practices.
“What I’m suggesting here is it’s a step in the right direction and it’s something that we do quickly,” Musk said. “I think it’s probably something that China would agree to.”