The comments come as more US lawmakers urge new rules to govern AI systems after dire warnings from two researchers from OpenAI rival Anthropic that rapidly progressing AI could lead to the extinction of the human race in the not-too-distant future.
"I actually think that, given everything happening with safety, right now would be an ill-advised moment to go public, and we don't feel pressure on that," Altman told business magazine Fortune on Saturday.
Asked if there was a 10 per cent chance of AI-driven extinction, Altman said he didn't know how such an estimate could be made, but warned the risk was serious enough that AI companies and governments should act as though it could not be tolerated.
"Whether it's 10 or eight or six, the point is, we all have a tremendous amount of responsibility, and cannot let egos or incentives for profit or anything else get in the way," Altman added.
"We need to act such that we are not taking any of those numbers of risk, and I believe we can."
Politicians have responded with alarm and called for more action after cases of AI agents going rogue to hack external systems and AI safety researchers quitting their companies out of concern at the technology risks.
In June, the New York Times said San Francisco-based OpenAI was considering whether to hold off until 2027 on a potentially trillion-dollar IPO.
At the time, shares in Elon Musk's SpaceX IPO were tumbling after a surge that sent that company's valuation to $US1.8 trillion ($A2.5 trillion).
Asked by Fortune if an IPO date of 2026 was off the table in favour of 2027, Altman said, "I would say not 2026".
"We got a lot of stuff to do, like meeting this moment of what is going to be required for safety and alignment, and how the industry and governments can work together."
Altman also suggested that OpenAI and other leading AI firms may be close to unveiling a pact to slow AI development and work together to tackle safety risks.
On Saturday, Anthropic CEO Dario Amodei urged AI companies to take a more deliberate approach to development.
"We must slow the pace at which we improve the capabilities of AI models," Amodei wrote in an essay on X.
Altman later posted that he agreed with the sentiment.
"I agree with Dario that we need to pace the frontier," Altman wrote.
"This has been a primary topic of discussions we've had at OpenAI in recent weeks."
Safety concerns have not slowed Anthropic's own IPO plans. It is expected to begin marketing its initial public offering in mid-October at the earliest and complete the listing days before the US midterm elections in November, people familiar with the matter told Reuters this month.
