Mark Zuckerberg rejects coordinated AI slowdown, backs company-led safety

Mark Zuckerberg said AI companies should decide their own safe development pace rather than slow down together. His stance sharpens the split in Silicon Valley over how to manage fast-advancing AI.Meta CEO Mark Zuckerberg has said each artificial intelligence company should decide the pace at which it develops its technology safely, distancing himself from calls by some rivals for a coordinated slowdown in advanced AI development.His comments, posted on social media, came after Anthropic CEO Dario Amodei proposed a three-step plan on AI safety, including coordination among companies and countries to slow development. The proposal quickly drew support from tech leaders including OpenAI's Sam Altman and Elon Musk.Zuckerberg's remarks come amid a wider debate among industry leaders, sparked by warnings from researchers about AI's potential threats to humanity and how to control the technology before it could escape human control.He said AI companies are already motivated to build their models safely and face "significant liability" to prevent them from causing harm. "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," he wrote on X. Without naming Anthropic or any other company, Zuckerberg said Meta had delayed the release of its Muse AI agent for several months to ensure its safety and security. "We didn't call for everyone else to do this before we would," he said.Zuckerberg also said companies should focus on building safe AI models instead of systems that can automatically improve themselves, a function known as recursive self-improvement. "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," he wrote. In sum, Zuckerberg argued that AI safety should be handled by individual companies taking responsibility for their own systems, even as debate continues over proposals for a broader, coordinated slowdown in AI development.With PTI Inputs- EndsPublished By: India Today Web Desk Published On: Sep 16, 2026 19:30 IST