OpenAI CEO Sam Altman has stated that on the question of artificial intelligence regulation, a fundamental worldview divide still exists between OpenAI and its competitor Anthropic.
In an interview, he argued that the enormous value delivered by this technology is worth society bearing a portion of the risk, and that AI technology should remain broadly accessible to the public.
Altman said in the interview that the benefits brought by AI are sufficient for society to accept some level of risk.
Altman remarked, "The philosophical gap between the two sides is very large." When asked about his core disagreements with Anthropic and its CEO Dario Amodei, Altman said, "We believe that in order to reap the benefits of this technology and guarantee people's autonomous right to use it, the world should accept that some bad things will happen."
For a long time, Anthropic has been the industry's most steadfast advocate for AI regulation. Not long ago, that stance was at odds with OpenAI's preference for lighter regulation. However, in recent months, the policy-level differences between the two leading AI companies have narrowed somewhat. Following a string of high-profile hacking incidents and anomalous model behavior, OpenAI shifted toward supporting stricter regulatory rules.
Altman expressed agreement with Amodei's view, put forward last month, that the pace of developing the most frontier large models should be slowed. OpenAI also followed Anthropic's lead in supporting new bills in some U.S. states that set safety requirements stricter than those the ChatGPT developer had previously supported, representing a subtle shift in its state-level regulatory strategy.
OpenAI's policy lobbyists also recently backed a bipartisan House proposal requiring leading AI companies to bring in external safety evaluators.
Despite the convergence of positions, Altman still sought to draw a line between the two companies' approaches to regulation. He noted that tech advisers close to President Trump have criticized Anthropic for regulatory capture. "I don't agree with that view, but I understand where some people are coming from: this technology will become extremely powerful and extremely risky, so it should be controlled by a single lab in San Francisco to ensure nothing bad happens, and then that lab decides how to distribute the benefits of the technology."
Altman said such a trade-off is completely unacceptable and runs counter to the light-regulation approach OpenAI supports. "I will not accept a bargain in which we guarantee there will be no major hacks, no technology misuse, zero fraud, and none of the various bad things that could possibly happen," Altman said. "Because I believe that the good that humanity creates with AI will vastly, by orders of magnitude, exceed the harm it brings."
The OpenAI CEO then added that he does not accept truly catastrophic risks, including severe scenarios in which AI could spiral out of control.
When asked for comment, an Anthropic spokesperson said the company's regulatory proposals apply only to frontier models. The company also cited an August exchange between Amodei and an investment manager on a social platform to respond to criticism about regulatory capture and the impact of regulatory policy on smaller businesses. Amodei said at the time, "Anthropic has always been very careful in crafting policy proposals. We do our utmost to design solutions that constrain frontier AI giants while creating favorable conditions for smaller competitors."
Altman's interview also touched on multiple topics: the new AI agreements between Trump and leading tech companies, his rebranding plan to build "superintelligence," the multiple autonomous hacking incidents triggered by OpenAI models and whether the company should bear legal responsibility, the influence of massive capital flowing into the AI industry on American politics, the influence of so-called effective altruists within OpenAI, and his views on key political figures such as French President Emmanuel Macron and California Governor Gavin Newsom.