The Escalating Effort to Slow AI

The momentum to slow the development of artificial-intelligence models now seems to mirror the speed at which AI capabilities were racing ahead.

The AI backlash that began as a turn against data centers by people across the political spectrum shifted in the last few days to Terminator-level fears.

First, Anthropic researcher Jacob Coxon resigned over his concern that AI companies are speeding toward a time when they won’t be able to control what they’re building, the Journal reported last week.

That same Journal story then quoted Evan Hubinger, a top Anthropic scientist, echoing those concerns on X: “We really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”

The alarm spread far and wide. In another Journal story, Rep. Josh Gottheimer told of being questioned about what’s happening with AI at his daughter’s field hockey game.

On Saturday, Anthropic CEO Dario Amodei, citing the safety issues, wrote on his personal blog that “we must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain.”

“I agree with Dario that we need to pace the frontier,” rival Sam Altman, the OpenAI CEO, wrote on X. He added that “committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We’ll have more to share soon.” (News Corp., owner of The Wall Street Journal, has a content-licensing partnership with OpenAI.)

SpaceX CEO Elon Musk wrote on X: “Dario is right.”

But Anthropic has raised the idea of slowing development before. My Journal colleague Tim Higgins this weekend spelled out the moral conflict gripping the company, noting the money at stake and the race not only among companies but also among countries.

Then he had this key point: “Yet, as thoughtful as Anthropic’s leaders are, they have failed to understand the only thing they can control is how they act. They can’t stop rivals from destroying the world, but they can decide they don’t want to be part of that.”

David Sacks, an adviser to President Trump, struck a similar note on X about Anthropic and OpenAI, writing that “if the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone else’s permission.”

Will a pause or slowdown actually happen? And what are CEOs to make of all this? The many chief executives we talk with at the Leadership Institute are all focused on making their significant investments in AI pay off. Their biggest challenge remains getting workers on board. Many employees already were worried AI would take their job. Will a fear that AI could also kill them add to chief executives’ challenge?

These major questions make it a perfect time for the Leadership Institute’s WSJ Technology Council Summit, which starts today in New York City. CEO Brief will bring you highlights this week.

Erle NortonManaging Editor, C-Suite Content

More from the Journal on the fallout: How the Clash Between Money and Safety Created a Monumental Crisis for AI

CEO Brief is written and edited by a team from The WSJ Leadership Institute: Alan Murray, president; Erle Norton, C-Suite Content managing editor; and Lila MacLellan, CEO Brief editor.

The Leadership Institute aims to redefine executive leadership development and peer networking by leveraging the unparalleled trust and insights of Dow Jones and The Wall Street Journal to create exclusive, transformative experiences for industry leaders that help them navigate the ever-changing landscape. To learn more, read on here.