The chief executive of OpenAI, the company behind the generative AI platform ChatGPT, Sam Altman testified before Congress recently, calling regulation of “increasingly powerful” AI necessary.

He also warned the technology could go “quite wrong” if safety standards are not sufficiently developed.

“We think that regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models,” Altman stated in prepared remarks. “For a very new technology we need a new framework.”

He called for a new agency with the power to license “any effort above a certain scale of capabilities” and which could take that license away and ensure compliance with safety standards.

Altman also conceded that AI was likely to make some jobs obsolete, but said that impact would be balanced by new, better jobs that the technology would create.

“GPT 4 and tools like it are good at doing tasks, not jobs,” he said. “GPT 4 will, I think, entirely automate away some jobs and it will create new ones that we believe will be much better.”

Among the other tech executives who testified on Capitol Hill were Christina Montgomery, chief privacy and trust officer at IBM, who called for precise regulation of AI in use cases, as opposed to regulation targeting development of the technology itself.

Afif Khoury, CEO, SOCi, a social data and intelligence technology startup, points out The narrative around AI hasn’t yet crystallized.

“Lawmakers are grappling with the implications of this technology, and Altman’s testimony seemed primarily focused on establishing a tone and tenor for how industry players, such as OpenAI, IBM, or Microsoft, might collaborate with regulators to create common-sense legislation,” he says.

He added that unlike the more heated exchanges seen with Mark Zuckerberg and others, the discourse this time around isn’t yet politicized, meaning regulations might be more predicated on genuine utility, consumer safety and innovation.

Khoury says Altman’s call for extensive regulation, including a new government agency responsible for licensing AI models, was somewhat surprising, but may be savvy for several reasons.