Sam Altman, chief executive of ChatGPT maker OpenAI, said yesterday that the public should trust AI companies to keep the technology safe, while accepting that people were right to be concerned about its risks.
“The world should trust that we are going to do the right thing because it’s the right thing and we feel the magnitude of this,” Altman said at Dreamforce, the annual gathering of Salesforce clients in San Francisco.
“It doesn’t take as much imagination as it used to for [us] to imagine how this could go wrong,” he said. “I think the world is right to be afraid of this.”
Altman said AI companies were capable of regulating themselves. “We will get it right,” he said, adding that if firms could not keep “alignment and safety way ahead of capabilities” they would “slow down or stop”.
The remarks were his first public comments since a post went viral last week from a researcher who quit Anthropic, claiming AI could “kill all humans” by the end of the decade if left unchecked.
Industry leaders back self-regulation
Nvidia chief executive Jensen Huang, speaking at the same conference, said AI firms should decide whether new versions of the technology are released, rather than outside forces.
“We don’t need new laws or regulations,” he said, adding there should not be a “false choice” between the speed of innovation and the safety of AI products. Huang argued that safety is an “engineering problem” and that companies should “take a pause” if they lose confidence in a product.
After Altman’s comments, Meta chief executive Mark Zuckerberg wrote on X that every AI firm had the ability and the incentive “to take its own actions” towards safety.
“Any lab that doesn’t focus on alignment will fall behind,” he wrote. “Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well.”
The recent warnings from inside the industry prompted Anthropic chief executive Dario Amodei to call for the pace of AI development to slow and for governments to regulate the sector. His call was applauded by Altman, Google DeepMind co-founder Demis Hassabis and Elon Musk.
Speaking at Dreamforce yesterday, Amodei said Anthropic was now in “a dialogue with the rest of the industry” about committing to better safety standards and checks on AI tools and development.
OpenAI executive Chris Lehane said last week that the company was working with other labs, including Anthropic and Google DeepMind, “to advance frontier AI standards, building a voluntary effort now, with or without government support.”
Critics question industry oversight
Critics say companies cannot be trusted to police themselves. Senator Bernie Sanders said decisions about AI had been left to “a handful of the richest people in the world”, while former White House strategist Steve Bannon said the public could not trust tech oligarchs to regulate themselves.
Jack Clark, an Anthropic co-founder and executive, told the BBC on Monday that leaving AI as a “totally unregulated industry” was “rolling dice with immense risks”.
Yoshua Bengio, one of the pioneers of modern AI, wrote on X that “ambitious efforts outside the for-profit sector” were needed to “avert the worst risks from AI”.
President Donald Trump has rejected calls for more guardrails, describing fears about AI safety as a “hoax” and saying the only guardrail needed was a “strong and smart” president.
New US legislation looks unlikely soon. In an interview with Axios published on 3 September, Altman said he had expected Congress to create a “basic framework” for advanced AI after his 2023 testimony, suggesting lawmakers had struggled to regulate without “slow[ing] down innovation”.
Congress returned from its summer break on Monday, but there is no consensus in either party on regulation, and the House is due to enter recess again at the end of this week ahead of the midterm elections in early November.
In the UK, Business Matters reported last week that Anthropic released its Claude Mythos 5.1 model without submitting it to the AI Security Institute for testing.
