Is Big Tech's AI slowdown proposal a genuine safety measure or competitive gatekeeping?

When executives from OpenAI, Anthropic, Google DeepMind, and SpaceX recently signaled support for slowing AI research, observers immediately disagreed about the true intent. The leaders proposed embedding external auditors, tightening oversight of domestic labs, and achieving global coordination on development pace. Yet skeptics contended the real goal was crushing emerging rivals, undermining open-source work, and dodging binding legal constraints — some called it a textbook industry "cartel."

The complicated reality#

The landscape is messier than either interpretation suggests, according to voices spanning the sector. The three-part framework, detailed in an essay by Anthropic's Dario Amodei, echoes proposals AI safety researchers have championed for years. It could serve as a substitute for actual government oversight—yet under the Trump administration, formal regulation appears unlikely anyway. Experts contend that while the direction matters, the industry's own leaders may not be the right voices steering the conversation.

"The industry as a whole needs new champions," said Nick Reese, an adjunct professor at New York University and former emerging tech policy director for the Department of Homeland Security. The problem, he explained, isn't that these executives lack expertise; it's that nobody has articulated a credible vision of where AI is actually headed.

Building pressure from within#

Cracks appeared in industry unity when Jacob Coxon, a researcher at Anthropic, quit and posted a public letter warning that both OpenAI and Anthropic were "racing straight to self-improving superintelligence and gambling with our lives." The message resonated widely—attracting over 170 million views on X and widespread media coverage—though safety advocates note the concerns weren't novel.

Daniel Kokotajlo, who led research at OpenAI before founding the AI Futures Project, pushed back against the notion that company executives deserved credit. "People outside the companies have been calling for this for years," he said. After months of researchers and employees voicing alarm—including over 1,000 lab workers signing a public letter last July—the CEOs were "bowing to that pressure and also claiming credit for it."

Welcome but fragile commitment#

Many safety experts view the announcement positively, if cautiously. Apollo Research's CEO called it "one of the best things for safety in a long time " Buck Shlegeris at Redwood Research said he was "cautiously optimistic," while others voiced qualified support. Yet all stressed that converting informal promises into binding, verifiable agreements remained the core challenge.