OpenAI CEO Sam Altman has argued that humans must retain ultimate authority over the future as superintelligence develops, saying that preserving human control should impose boundaries on AI progress.
In a statement published on X, Altman rejected a future in which people lose their decision-making power and become limited to activities that artificial intelligence systems effectively permit or reward. According to ANI News, he said human oversight must remain a fundamental condition of technological development.
Human authority comes first
Altman's latest position centers on a distinction between developing increasingly powerful AI and allowing those systems to determine humanity's direction. He argues that technological capability alone should not establish who makes decisions about the future.
He maintains that human control must act as a constraint on the development of superintelligence. This position challenges approaches that could give advanced AI systems greater authority while leaving people with diminishing influence over important decisions.
The comments contribute to an ongoing technology-industry debate about the speed of AI development and the safeguards needed as systems become more capable.
Altman backs slower development
The statement follows a discussion in September, when Dario Amodei, chief executive of AI company Anthropic, called for the industry to moderate its development pace. Amodei warned that improvements in AI capabilities could move faster than researchers' ability to understand and control the technology.
Altman supported that proposal at the time, expressing agreement with Amodei's position on managing the pace of frontier AI development. He also backed allowing independent evaluators to examine AI systems, providing an additional layer of scrutiny over their safety.
Warning against concentrated power
Altman also addressed human oversight at the United Nations Security Council in September. He argued that developers should not train AI systems without making a convincing case that the technology would remain under human control.
He additionally cautioned against allowing the influence of advanced AI to become concentrated among a limited group of companies or national governments. His concern extends beyond individual model safety to the distribution of decision-making power as AI capabilities expand.
These positions establish two connected priorities: ensuring that AI systems remain controllable and preventing a small number of powerful actors from dominating the technology's development and use.
Safety commitments face scrutiny
The debate is unfolding as leading companies continue releasing more advanced AI models. A recent report by The Washington Post noted that OpenAI and Anthropic had continued introducing new systems despite public appeals for slower development.
That tension raises questions about how companies will translate their stated safety principles into actual decisions about research, training and product releases.
OpenAI has also expanded its cybersecurity work through a partnership with security company Cathedral. The collaboration aims to identify vulnerabilities and improve protections against emerging threats connected to AI.
Why it matters
Altman's remarks put human decision-making at the center of the superintelligence debate. Rather than treating safety solely as a technical problem, his position emphasizes who retains authority when AI systems become increasingly capable.
The immediate challenge is turning that principle into enforceable development practices. Independent evaluations, meaningful human oversight and safeguards against concentrated power are among the issues raised by his previous statements.
The next test will be whether the industry's public commitment to human control influences how leading developers build, evaluate and release their most advanced systems. For OpenAI, Altman's latest comments make clear that technological progress should not come at the expense of humanity's authority over its own future.