“We will get it right,” he said.
Alignment is the field concerned with making AI systems behave consistently with human values and ethical principles. That sounds straightforward until people have to agree on the values, encode them into software and ensure the software continues following them as its abilities expand.
Meta chief executive Mark Zuckerberg supported the industry-led approach after Altman’s appearance. Writing on X, Zuckerberg said every AI company had both the ability and the incentive “to take its own actions” on safety.
“Any lab that doesn’t focus on alignment will fall behind,” he wrote. He added that companies could face substantial liability if their models caused harm, giving them another reason to prevent dangerous outcomes.
That argument assumes commercial incentives, legal exposure and responsible engineering will produce enough caution. Critics respond that companies also face intense pressure to release stronger products before their rivals do, a small complication in the trust-based model.
Why US regulation remains unlikely
Altman has previously expressed frustration with the lack of federal action. In an Axios interview published on 3 September, he said he expected Congress to create a “basic framework” for advanced AI after he testified before lawmakers in 2023. No such framework has emerged.
He suggested legislators have struggled to understand a rapidly changing technology and regulate it without slowing innovation. Congress is unlikely to pass major AI safety legislation soon, despite lawmakers returning from their summer break on Monday.
Neither major US party has reached an internal consensus on what regulation should involve. The House of Representatives is also due to begin another recess at the end of the week, ahead of the midterm elections in early November. Artificial intelligence may be moving quickly, but the congressional calendar remains committed to its own pace.
President Donald Trump has rejected demands for stronger safeguards, describing concerns about AI safety as a “hoax.” He said the only guardrail the technology required was a “strong and smart” president.
Opposition to industry control also crosses conventional political lines. Senator Bernie Sanders has said decisions about AI have been left to “a handful of the richest people in the world.” Steve Bannon, Trump’s former White House strategist, has similarly argued that the public cannot trust technology oligarchs to regulate themselves.
Nvidia says safety is an engineering issue
Nvidia chief executive Jensen Huang also defended corporate control during Tuesday’s Salesforce conference. He said AI companies, rather than outside authorities, should decide whether new systems are ready for release.
“We don’t need new laws or regulations,” Huang said. He rejected what he called a “false choice” between rapid innovation and product safety.
Huang described safety as an “engineering problem” and said companies should “take a pause” if they lose confidence in a product. The position carries considerable weight because Nvidia supplies much of the hardware driving the AI boom. Surging demand for its chips has made it the world’s most valuable company by market capitalization and sharply increased its profits.
For developers, treating safety as an engineering matter places responsibility with the people who understand the systems most closely. For skeptics, it also leaves release decisions with companies that earn enormous sums when development continues.
Industry warnings have raised the stakes
Altman’s remarks were his first public comments since a post by a researcher who resigned from AI company Anthropic went viral last week. The researcher claimed that unchecked artificial intelligence could “kill all humans” by the end of the decade.
Several AI executives and specialists said they agreed with that assessment. They did not explain in detail how they calculated the risk or precisely how AI might eliminate humanity, leaving the public with an alarming conclusion and rather limited supporting material.
The warning increased scrutiny of the industry. Anthropic chief executive Dario Amodei called for AI development to slow and urged governments to regulate the sector. Altman, Google DeepMind co-founder Demis Hassabis and Elon Musk applauded his intervention.
By Tuesday, however, several prominent executives were again emphasizing voluntary corporate controls rather than government oversight. The industry’s message therefore remains somewhat divided: the danger may be extreme, development may need to slow, but the companies responsible would still prefer to decide when that should happen.
Yoshua Bengio, one of the pioneers of modern artificial intelligence, argued that corporate action would not be enough. In a post on X, he called for “ambitious efforts outside the for-profit sector” to “avert the worst risks from AI.”
The dispute is no longer about whether advanced AI presents serious risks. Many of the people building it now say that it does. The unresolved question is whether their concern should be treated as proof that they can manage the danger, or evidence that someone else should be watching them.
