The debate over who should control frontier artificial intelligence has moved beyond researchers and regulators. The executives building the technology are increasingly questioning whether companies like their own should have so much power in the first place.
Anthropic CEO Dario Amodei has repeatedly said he is uncomfortable with a handful of private companies making decisions that could reshape society. In a CBS interview, he agreed that nobody had elected him or OpenAI CEO Sam Altman to make those choices and argued that this is precisely why AI needs thoughtful regulation.
That position has become even more significant this week.
Amodei is now urging the industry to slow the frontier enough for safety systems to catch up, warning that increasingly capable AI agents could create far more serious cyber, biological and control risks within months rather than years.
OpenAI Is Moving Closer to Anthropic’s Position
Amodei is no longer isolated.
OpenAI said on Sept. 9 that it supports mandatory national AI safety requirements, including capability-based rules and independent safety assessments. The company argued that neither corporations nor governments can manage advanced AI risks alone.
Sam Altman then hinted that the leading AI companies could soon announce a broader industry pact.
Asked whether OpenAI, Anthropic, Elon Musk’s AI operations and Google DeepMind could jointly coordinate on safety, Altman said he believes such cooperation will happen, while declining to disclose private discussions.
That follows growing internal pressure across the industry.
A former Anthropic and OpenAI researcher recently resigned after warning that the labs were racing toward self-improving systems without adequate safeguards, a development covered in our report on the latest AI safety resignations.
The debate is especially striking because these companies are simultaneously heading toward enormous valuations. Anthropic and OpenAI have both been preparing for potential public listings, while Coinpaper has tracked their combined pre-IPO market ambitions.
Trump Rejects the Case for Slowing Down
Washington is moving in the opposite direction.
President Donald Trump has rejected calls from technology executives and Democrats to slow frontier AI development, arguing that the United States cannot risk surrendering technological leadership to China.
That creates an unusual split.
The companies with the strongest financial incentive to keep scaling are increasingly saying some form of restraint is necessary, while the White House worries that restraint itself could become a national-security risk.
The tension is even more complicated for Anthropic. Earlier this year, the Trump administration clashed with the company over restrictions on military uses of Claude. The Pentagon still considers Anthropic a supply-chain risk despite more conciliatory signals elsewhere in the administration.
Elon Musk also illustrates how messy the emerging coalition is. He has repeatedly warned about advanced AI risk, yet recently dismissed some of the latest Anthropic-linked safety warnings as a “psyop.”
Be the first to comment