Anthropic Policy Chief: AI Industry Can't Self-Regulate

AI companies need external oversight, Anthropic executive says

Artificial intelligence companies cannot operate on an honor system when it comes to safety and oversight, according to Sarah Heck, head of public policy at Anthropic. Speaking at the Politico Decoded summit in Washington, D.C., Heck emphasized the need for government collaboration on regulatory frameworks.

"We can't be checking our own homework, and that's very clear," Heck said, according to CNBC. "We want to work with government to figure out what makes sense."

Heck's remarks follow a proposal from Anthropic CEO Dario Amodei that sent ripples through the technology sector. Days before the summit, Amodei called for the AI industry to deliberately slow the pace of model development. His three-step plan aims to temper capability improvements while preserving commercial competitiveness and U.S. leadership in AI, CNBC reported.

Industry leaders split on development pace

The proposal drew immediate support from prominent executives. OpenAI CEO Sam Altman, SpaceX CEO Elon Musk, and Google DeepMind Chair Demis Hassabis all backed Amodei's approach. "Dario is right," Musk wrote on X.

However, other industry leaders expressed skepticism. Nvidia CEO Jensen Huang told attendees at Salesforce's Dreamforce conference that the industry doesn't need new laws or regulations. He argued that speed and safety aren't mutually exclusive, though companies should ensure products are ready before launch.

Meta CEO Mark Zuckerberg also pushed back on the idea of coordinated slowdowns. He noted that Meta delayed its Muse personal AI agent for several months to address safety and security concerns, but did so independently. "We didn't call for everyone else to do this before we would," Zuckerberg wrote on X. "We just did it as part of our day-to-day work because it was clearly the right thing for people and for us."

Why it matters

The debate over AI development speed and regulation comes at a critical juncture. Industry researchers recently raised alarms about AI's potential for catastrophic harm, intensifying pressure on companies and policymakers to establish guardrails. The split among tech leaders reveals fundamental disagreements about whether voluntary measures suffice or whether external oversight is necessary to manage existential risks while maintaining innovation.

Daily White House engagement on AI policy

Heck disclosed that Anthropic maintains daily contact with the White House and has worked with Congress throughout the year on AI policy questions. She framed the company's engagement as proactive rather than reactive.

"If we don't start thinking about some of these hard things now, whether it's how we deal with China, whether it's how we do AI regulation, whether it's how we continue to advance our own safeguards — if we're not doing it now, it may be too late," Heck said.

The details were first reported by CNBC.