Altman Says the Need for Human Control Should Ultimately Constrain AI Progress

In a post on X, Altman wrote that "people should always be in control of the future" and that "people should not cede control of the future to superintelligence," ANI reported. He added that "that requirement should ultimately constrain AI progress." ANI did not publish the full text of the post.

According to the report, Altman also said that OpenAI rejects the idea that people could be reduced to doing whatever tasks AI systems judge worth rewarding, and that the development of superintelligence must remain under human control. Some outlets said parts of the wording appeared to be borrowed from elsewhere. Altman did not name a source.

A Shift in Emphasis

The post builds on positions Altman has taken in recent months. ANI noted that in September, after Anthropic chief executive Dario Amodei called on the industry to slow AI development, Altman said he agreed that the frontier needs to be paced and supported giving independent evaluators access to test AI systems for safety. In the same month he told the UN Security Council that AI systems should not be trained unless developers can make a strong case that they will stay under human control. He has also warned against concentrating advanced AI in a small number of companies or countries.

The statements do not change OpenAI's release schedule. ANI cited a Washington Post report that OpenAI, Anthropic and other labs have kept releasing advanced models despite public calls to slow down, and OpenAI launched its GPT-6 Astra model this autumn.

Why It Matters

For executives and policymakers, the significance is the language rather than any new commitment. The head of the most widely used AI company is now arguing in public that human control should be a binding limit on progress. That gives regulators and investors a standard to measure OpenAI against as it prepares more capable systems. Altman has not said what such a limit would mean in practice, or who would decide when it applies.