All of a sudden, all the AI mavens are aligned. Sam Altman, head of OpenAI, said this week that his company welcomes “a federal framework that sets consistent safety requirements for frontier AI.” His worry? That the industry might lose “control of the future to AI.” Anthropic’s Dario Amodei agrees. “Left unchecked,” he wrote in a recent essay, AI “could outrun our ability to understand and control these systems.” Even Elon Musk concurs, noting that he has been “sounding the alarm” for years.
Color us skeptical.
It is, of course, entirely possible that researchers at OpenAI, Anthropic, and Grok are aware of troubling developments to which we are not privy. But we remain suspicious of early movers who call for government regulation. Sam Altman insists that “no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.” One must wonder which of those terms is the most important. Is it “recklessness”? Or is it “competitive pressure”? The current leaders in the AI realm are in a highly advantageous position and would stand to improve that position were the government to entrench them as virtual utilities.
That advantage notwithstanding, the practical case for a slowdown is weak. Handing control to the federal government merely moves the risk elsewhere. Are we really to believe that Washington, D.C., has more expertise than the companies that have developed the technology? And, if so, ought we to consider it a more responsible actor? (Covid, after all, very likely came out of a government-run lab.) The Trump administration is vocally against a slowdown — much more so, in fact, than the companies they’d be regulating. If OpenAI, Anthropic, and others want one, shouldn’t they just . . . do it?
Dario Amodei has suggested that such a halt would be “very difficult” and admitted, “Honestly, I don’t know if it’s possible.” This is, in part, because there is no way of forcing China and our other global adversaries to follow suit. That being so, we would favor erring on the side of caution — which, in this circumstance, means ensuring that the United States retains its lead in this contest. Given the enthusiasm with which China rips off American intellectual property, there is no reason for anyone to believe that its leadership would respect an international slowdown pact or that it would hesitate to press its advantage once obtained. International competition is red in tooth and claw, and the best form of defense is offense. We cannot afford to fall behind.
Nobody of a responsible disposition wishes to be reckless. But, by the same token, we ought to avoid panic — especially when that panic lines up so neatly with the commercial ambitions of its progenitors. At some point, there may well be a case for a different approach to AI, but, for now, we’d counsel keeping calm and carrying on.