OpenAI CEO Sam Altman, in a Fortune interview, said that even a 10 percent probability of the technology causing humanity’s extinction is unacceptable. He added that under the current security conditions, taking the company public in 2026 would be the wrong move and that he feels no external pressure to do otherwise. Altman stressed that risk levels—whether 10 percent, 8 percent or 6 percent—demand that no one let ego or profit override the responsibility to prevent such an outcome.
The warning has been echoed by researchers at competitor Anthropic, who recently cautioned that the rapid pace of AI advancement could bring about humanity’s end. U.S. lawmakers from both parties have called for new regulations after incidents involving autonomous agents that targeted external systems and the departure of security researchers from the sector. A June New York Times piece noted that San Francisco‑based OpenAI management is evaluating whether a trillion‑dollar IPO is viable, a decision that followed a sharp decline in shares after SpaceX’s $1.8 trillion valuation announcement.
Altman said 2026 will not be the year for an IPO, citing the need to meet safety and compliance standards and to work closely with governments. Fortune reported that, based on his remarks, OpenAI and other industry leaders might soon announce an agreement to slow AI development and collaborate on security risks. Anthropic CEO Dario Amodei, in a recent X post, urged firms to slow the speed of model improvement, a sentiment that Altman echoed on the same platform.
