Anthropic’s top executive calls for government measures aimed at slowing AI development (and the advances of rival firms). Has such an approach ever borne fruit?
Mobile devices made everyday life easier. So did the web.
Artificial intelligence could make it even better.
Yet many people are uneasy.
“Stop AI or we’re all doomed!” chant the demonstrators.
Some fear that AI might decide humans stand in its way and confine or even harm us. Not the most likely scenario, according to AI experts, but within the realm of possibility.
The head of Anthropic recently warned that AI is advancing beyond our capacity to manage it. He and other AI leaders advocate slowing its pace.
His idea includes placing independent “evaluators” inside the company to gauge risk. He also calls for international cooperation and national rules that would curb the pace of innovation for others as well.
Has that ever produced lasting results? I doubt it.
“The anti-AI movement repeats the same errors as the anti-fossil-fuel campaign,” says author Alex Epstein in my new video. “They deny the enormous benefits of something, blow hazards out of proportion, and ignore practical ways to address them.”
Street demonstrations target the water used to cool data centers powering AI. Yet data centers consume only a sliver of U.S. water (about 0.04 percent in 2024).
There’s also the option of using recycled water.
Some worry that the demand for electricity from data centers will push up energy prices.
Epstein argues that AI can be run in a way that keeps electricity costs steady, but achieving that would require reversing policies that have restricted electricity generation.
Such steps would include shutting down nuclear plants and mandating greater reliance on renewable power.
“With AI,” Epstein contends, “we’ve made things harder than necessary. Adopt policies that facilitate expanding electricity capacity and enable AI to establish its own private grids.”.
Why not pursue this?
“There aren’t no concerns,” Epstein concedes, “but data centers can become valuable community assets that basically operate like cash machines… People have no idea what they’re sacrificing… A lot of the work you regard as tedious—like grading half the time for teachers or compiling forms for nurses—could be delegated to AI, making teachers more effective and nurses more efficient.”
AI will displace some jobs as well. Yet forecasts that automation will wipe out vast swathes of employment have repeatedly proved wrong.
We were told automation would yield 15 percent unemployment. Automation arrived, and employment actually grew.
Adam Thierer, author of Permissionless Innovation, argues, “History shows that technology has often been a friend to humanity and used for great good. That will include AI…if we permit it.”
Will we permit it?
Senator Bernie Sanders (I–Vt.) has proposed government ownership to the tune of 50 percent.
“If the government owned half,” Thierer says, “we’d likely stifle the development of high-quality AI. Governments don’t have a strong track record of forging cutting-edge tech that serves the public well. I don’t want the U.S. Postal Service managing our internet or our communications networks, and I don’t want Bernie Sanders…controlling our AI robotics.”
Even if American firms slow AI, progress won’t halt entirely.
“If we misstep,” Thierer notes, “the Chinese are ready to step in… They’re investing heavily to claim global leadership. Their choices will be embedded in their systems: censorship, surveillance, and control.”
China has declared that all AI must conform to “core socialist values.”
“Whoever controls AI wins,” asserts President Donald Trump. He rejects any attempt to slow U.S. AI innovation. And there’s merit to that view, in his frame.
AI will shape how billions interpret the world. We should root for technologies that advance freedom to prevail.
COPYRIGHT 2026 BY JFS PRODUCTIONS INC.
