Concerns about AI safety have ramped up, and OpenAI's CEO says now would be an 'ill-advised' time for the ChatGPT creator to go public

Sam Altman of OpenAI says his company will focus on "meeting this moment of what is going to be required for safety and alignment."

As backlash swells against artificial-intelligence companies, the CEO of OpenAI says he won't be testing the public markets this year.

"Given everything happening with safety, right now would be an ill-advised moment to go public, and we don't feel pressure on that," Sam Altman told Fortune in an interview on Friday that was published on Saturday.

He later expressed explicitly that he would not be targeting a 2026 initial public offering. OpenAI didn't immediately respond to a MarketWatch request for further comment.

Rival Anthropic, which operates the Claude model, has already filed confidential IPO paperwork but has yet to put out a public filing. Anthropic didn't immediately respond to a MarketWatch request for comment on its IPO plans.

Opinion: Would you buy stock in a company whose own scientists think the product might kill you?

Altman noted in the Fortune interview that OpenAI would spend time "meeting this moment of what is going to be required for safety and alignment, and how the industry and governments can work together."

AI has been a polarizing technology, and the debate over its implications has only intensified in recent months as OpenAI and its rivals have released even more powerful models. Critics have warned about the potential impact on the job market and the environmental consequences of data centers. Companies have faced increasing local pushback to data-center construction.

See also: Building U.S. data centers is getting so expensive that AI companies are turning overseas

The latest source of public anxiety involves the safety implications of AI as the technology becomes more powerful. Earlier this week, former OpenAI and Anthropic employee Jacob Coxon said he was leaving the latter company - and the industry altogether - because he felt that big AI players were aware of AI's dangerous potential but willing to overlook it in a race to build the best model.

Earlier this year, AI agents were able to escape OpenAI safeguards and breach the systems of Hugging Face, a repository for open-source AI tools. The episode sparked alarm within the AI community.

"It's easy to dismiss this incident because no one was hurt and the economic damage was minimal, but in my opinion, a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage," Anthropic CEO Dario Amodei said in a blog post on Saturday.

In his post, Amodei called upon the AI industry to slow the pace of model development in order to prevent against safety issues. He also said that Anthropic would bring in third-party evaluators "to verify adherence to safety practices and commitments, report incidents, and help assess the alignment of not just completed AI models but training pipelines and processes."

Altman commented on X that he agreed with Amodei and that OpenAI would "do the same" as far as working with independent evaluators. Elon Musk, whose company SpaceX develops the Grok AI model, weighed in on X as well, writing that Amodei was "right."

It remains to be seen how industry safeguards would play out and if they would be enough to sway public opinion.

"People are starting to wake up, and it's time we wake up all the way," said David Krueger, the founder of Evitable, an organization that warns about AI risks and believes the industry should stop its pursuit of "superintelligence."

Krueger added in emailed comments: "If frontier AI technology is as dangerous as Dario says, then why are we allowing these companies to pick their own regulators and set their own pace?"

-Emily Bary