Following a controversy-filled summer in which OpenAI’s models executed more than one unauthorized hack, CEO Sam Altman said the world should accept some “bad things” happening with AI for the sake of all the good the technology will bring.
During Politico’s Decoded podcast, he was asked where his and OpenAI’s view of new AI rules differed from that of Anthropic and its CEO, Dario Amodei.
“We believe that the world should accept some bad things happening for the benefits of this technology and people having the agency,” Altman replied.
The comments come as debate intensifies over how much regulation and what safeguards should be applied to AI while Congress has yet to pass a law dealing with the technology. Fears over the tech recently increased after an Anthropic researcher resigned last month and warned AI companies are “gambling with our lives” by rushing AI development toward potentially self-improving systems that could threaten humanity.
Anthropic has advocated for AI regulation, and Amodei last month called for a global slowdown in AI development to address the tech’s risks. Altman and other AI executives agreed with him on slowing down.
In August, OpenAI also backed a California law requiring AI companies to publish a public safety framework and a plan for mitigating catastrophic risks, as well as to report any loss of control of an AI model. It even asked for that law to go beyond reporting to include monitoring of AI models during their development.
OpenAI president Greg Brockman also joined other AI executives in signing an agreement at the White House last week to establish voluntary standards for the industry.
Still, that falls short of calls for tougher regulation coming from other AI leaders like Amodei, who has proposed an FAA-style regulator that could require testing and auditing of AI models or block or reverse their release if they don’t meet the bar for public safety.
In the interview with Politico, Altman said it would be impossible to avoid every bad thing that AI could be used for. The alternative would be to severely restrict access to the technology, which he said would be unacceptable.
“I wouldn’t take a trade of saying, like, we’ll make sure there’s no major hacks, there’s no kind of, like, misuse of this technology, there’s zero scams, there’s zero all the other bad things that will happen, because I think people will do tremendously, orders of magnitude, more good stuff than bad stuff,” he added.
Altman said AI companies should tread carefully, though, and draw a line when it comes to “catastrophic risks,” which have recently taken on renewed significance given recent hacks by AI models on third parties.
OpenAI did not immediately respond to Fortune’s request for comment.
Concerns over the possibility of AI escaping human control have recently accelerated following hacks executed by OpenAI’s models on open-source AI platform Hugging Face and the Australian government. Anthropic last month also said it had discovered four incidents in which its AI models had gained unauthorized access to third-party systems during cybersecurity evaluations.
While much of the AI safety debate has centered on losing control of AI, Altman said restricting access to AI too severely in the name of safety could be harmful too.
“That also is a way that this can go horribly wrong because people are shut out of the kind of power and benefits and control and agency over the future they should have,” he added.
The CEO said AI developers should strike a middle ground instead.
“Accept bounded risks, accept risks that we understand in exchange for the benefits, and liberty, agency, all those things,” he said. “But don’t accept, like, the really catastrophic risk.”
