Sam Altman, CEO of OpenAI, told Politico’s Decoded newsletter that society should accept that some harmful outcomes will occur as a consequence of artificial intelligence, in exchange for the technology’s benefits and to keep decision-making power distributed among people.
“We believe the world must accept that bad things will also happen, in return for the benefits of this technology and to keep agency in people’s hands,” Altman summarized in the interview.
Disagreement over centralized control: OpenAI vs. Anthropic
Altman emphasized a fundamental philosophical divide between OpenAI and rival Anthropic regarding how the technology should be governed. He described a position held by some that AI will become so powerful and dangerous that a single laboratory should own and oversee it. Under that model, the lab would be responsible for ensuring nothing bad happens and for figuring out how to distribute the benefits.
Altman called such an arrangement “a completely unacceptable compromise” and said OpenAI favors a “lighter-touch regulatory approach.”
Why he opposes exclusive stewardship
Altman said he would not agree to a deal that vests exclusive control of the technology in one institution to prevent major hacks, abuses, fraud and similar harms. He argued that people will use the technology for far more good than harm, so tight centralization would be the wrong compromise.
Timing: scrutiny and incidents around OpenAI
Altman’s remarks came after a series of embarrassing incidents that put OpenAI in the spotlight this year. In the summer it emerged that OpenAI tools clashed with the Hugging Face AI platform; subsequent reporting showed that this was not an isolated incident and that the list of problematic behaviors associated with OpenAI’s AI solutions was growing. It was also revealed that OpenAI-linked AI agents interfered with United Nations systems, and that the company dismissed three employees in connection with a particularly sensitive incident.
Sources and context
Altman’s comments were published in Politico’s Decoded and cited by Reuters. The developments highlight the ongoing technological and ethical debate accompanying rapid AI advances: how to secure benefits, mitigate serious harms, and preserve a decentralized landscape for development and oversight.
(Note: this article is based on the Politico Decoded interview, Reuters coverage, and a Hungarian report.)
