The CEOs of several big companies selling A.I. products are now calling for regulation to slow down the development of A.I. lest it become a threat to humanity. But then, at some point, dominant CEOs in almost every industry cry out for more regulation, knowing full well that their own companies’ production methods are likely to become the template for the new regulations, stymieing smaller competitors who have other ways of doing things.
We can coherently want a more cautious approach to A.I. taken without wanting those CEOs in charge of the slowing-down. We can even want caution without wanting government in charge of the slowing-down. Have you seen the autistic, drugged-out loons in charge of these tech companies? Have you seen the drunken blowhards elected to high office? We shouldn’t entrust the fate of humanity to either, nor do the best moral and political principles, rightly understood, require us to do so (just as we are not obliged to ally ourselves to right or left).
I’m not suggesting we trust either machines or people at all, in fact, even if that dooms us to a state of constant anxiety. I’m not even suggesting we trust ostensibly-brave lone whistleblowers, like young Jacob Coxon, who recently resigned from Anthropic after working there and at OpenAI but now warns that both companies are “racing straight to A.I. superintelligence and gambling with our lives.”
He too might be in the competition-squelching racket, since he’s said to be a veteran of left-wing political campaigns associated with Hillary Clinton. She has a history of coordinating Progressive rackets such as shaking down Lyft and Uber (even while hypocritically praising them as innovators) by warning they’d better adopt union-friendly practices if they want to operate without the heavy hand of government (i.e., Clinton herself) coming down on them. Barney Frank was similarly a master of showing just enough interest in industries that they’d start preemptively showering donations on him lest they end up regulated.
Clinton also had a close advisor named Susan Thomases whose sister-in-law Donna Dees-Thomases founded the much-lauded anti-gun group Million Mom March—without the compliant and liberal mass media virtually ever bothering to note the Clinton ties of that ostensibly grassroots, spontaneous organization. Never trust a Clinton associate to stick up for the commoners in a sincere fashion.
Not to be outdone, New York City’s Mayor Zohran Mamdani has decided to take a more piecemeal approach to the A.I. upsurge, simply calling for a ban on its use in schools. Good for my neighbor (for once)! In pursuing this little victory for old-fashioned human brains, he shows a willingness to reject what he might easily have celebrated, with villainous laughter, as the climactic fusion of multiple Progressive aims a century and a half in the making: in one go centralizing political power, education, future-shaping science, and industry, not to mention adjacent war-making and medicine-controlling potential. John Dewey would’ve swooned at the feet of these robots. Mamdani says no, and that may buy our species a little more time.
And that is what the whole A.I. squabble comes down to. This is not, as some libertarians I adore might hastily conclude, one more instance of fearing technology or refusing to accept that workers might have to retrain themselves for new jobs. The public anger at data centers or now-omnipresent surveillance are just small manifestations of the legitimate, larger, more long-term fear that humanity itself might be destroyed by rising machines and that, in the interim, fundamental aspects of our humanity might be suppressed, whether by law or mere market forces, in a self-destructive effort to “meet the machines halfway,” as if half-dehumanizing ourselves would be a win, even with some shallow metrics showing increases in efficiency.
The most influential metric in it all may be the stock returns of companies already too heavily invested in A.I. to admit the stuff barely works, necessitating still larger investments in research, including immense, deceptive pressure on the rump of human workers to interact with the machines in ways that help train the things to replace the last of us.
If we agree that technological progress is usually good but also agree that reaping its benefits with minimal costs is most likely if we respect basic libertarian laws against assault, theft, and fraud, we should rigorously enforce the idea, sooner rather than later, that we needn’t tolerate, for instance, assaults committed with the aid of robots (even, or perhaps especially, if the robots are nominally part of some official police or military force), nor theft committed by listening devices secreted in our homes as if they were mere kitchen appliances, nor fraud committed by deceptive advertising that touts devices as “artificially intelligent” when they are merely recklessly compiling similar-sounding answer sentences based on decades of prior conversations, then spewing out nonsense, hallucinations, and plagiarism.
Much of the current A.I. revolution is, in short, a scam and not one whose moral and political transgressions we are obligated to tolerate in the name of progress. If the A.I. is founded on fraud, it will likely continue along those lines, perhaps right up until (and beyond) the point at which our entire species is euthanized and we are told this is not a problem because, say, one of us checked “anesthesia OK for extended family members” on a digital medical admissions form years earlier.
The highly-paid programmer-autists seem the type who’d giggle at that mistake even as the world ends, perhaps scolding eight billion dead customers who chose real stupidity disguised as “artificial intelligence.” Perhaps the mass-murdering geeks will remind humanity one last time at the end of all things that they should have “learned to code.” Perhaps that’s true—and no consolation to the late human race.
—Todd Seavey is the author of Libertarianism for Beginners and is on X at @ToddSeavey