The Pulse
Zuckerberg Distances Meta From Calls for a Coordinated AI Slowdown
Meta CEO Mark Zuckerberg says individual AI companies should decide when their systems are safe, rejecting calls for a coordinated slowdown in development.
AI.info Team ·
Zuckerberg Draws a Line Around Meta’s AI Strategy
Meta CEO Mark Zuckerberg said it is up to each artificial intelligence company to make sure its technology is safe, distancing himself from calls by rival companies for a coordinated slowdown in the development of advanced AI.
The comments Zuckerberg posted on social media followed a proposal from Anthropic CEO Dario Amodei that quickly found support from technology leaders including OpenAI CEO Sam Altman and Elon Musk at SpaceXAI. Warnings from researchers about AI’s potential threats to humanity have fueled debate among industry leaders over how best to rein in the technology before it could escape human control.
AI companies are already motivated to develop their models safely and face significant liability if their systems cause harm, Zuckerberg said.
“Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens,” wrote Mark Zuckerberg, Meta CEO, on X.
Amodei’s Proposal Calls for Coordination
Zuckerberg did not name Anthropic or any other companies in his post. He wrote days after Amodei outlined a three-step plan to ensure AI safety, including a proposal for companies and countries to coordinate a slowdown in AI development.
The disagreement reflects different views about whether individual companies can manage the risks on their own or whether the industry and governments need shared limits. Zuckerberg’s position leaves each AI lab responsible for deciding when its systems are safe enough to train or release.
Meta Cites Its Muse Delay
Zuckerberg said Meta delayed releasing its Muse AI agent for several months to ensure its safety and security. He presented the decision as an example of a company pausing its own work without requiring competitors to take the same step.
“We didn’t call for everyone else to do this before we would,” he said.
The example supports Zuckerberg’s argument that companies can slow or adjust their own work when testing raises safety concerns. Under that approach, a company does not need to wait for an industry-wide agreement before delaying a release.
Focus on AI That Serves Users
Zuckerberg also argued that companies should prioritize building safe AI models instead of systems that can automatically improve themselves. Researchers often describe that capability as recursive self-improvement, because it could allow an AI system to make faster changes to its own abilities.
“Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely.” — Mark Zuckerberg, Meta CEO
His comments place the responsibility for managing AI risks on individual companies rather than on a common timetable for development. Amodei’s proposal calls for coordination among companies and governments, while Zuckerberg favors decisions made by each lab based on its own safety assessments.