The maxim “Once bitten, twice shy” never seems to apply in politics. We don’t learn. An explosion of concern over the hurtling advance of artificial intelligence seems likely to follow the same pattern as previous panics, which goes like this:
A big scare bursts upon the world; knowing little about it, government and other vested interests massively overreact; naysayers who warn against unnecessary or excessive measures or propose alternatives are excoriated; when, too late, they are proven right, they get no apology; those most responsible for the damage cover their tracks, muttering, “Never acknowledge; never explain.”
It happened with COVID. We locked down all social activity except left-wing race rioting, imposed intrusive, science-free mandates, ruined hundreds of thousands of businesses, wrecked the education of millions of children, and stoked depression and alienation. Six years later, we still have not recovered.
So, too, with climate alarmism. We borrowed trillions of dollars and misallocated them to inefficient technologies and programs, hobbled economic growth, crimped the American dream, abandoned millions of the world’s poorest people to die in undeveloped countries, taught generations of children to be scared, and created a society that regards humanity as a blight on the planet and fails to reproduce itself.
There are many other examples, but you get the picture.
Now, suddenly, we’re on the verge of a repeat with AI. If Anyone Builds It, Everyone Dies, the title of a book published last September, captures both the substance of the alleged existential threat and the tenor of the consequent alarm that burst into headlines and the public imagination last week.
Dario Amodei, head of Anthropic, called for “pacing the frontier,” by which he means slowing AI development because those building it are losing control and their machines could threaten humanity. He wants government safeguards put in place so frontier companies can work together without antitrust prosecution. OpenAI’s Sam Altman immediately agreed, as did others, including Elon Musk.
Former President Barack Obama said Democrats should develop an AI regulation policy. Sen. Jon Ossoff (D-GA), running for reelection in Georgia but also tipped as a presidential possibility, jumped on the bandwagon, and naturally, Sen. Bernie Sanders (I-VT) waits in the wings with legislation to stop data centers from being built and for a socialist federal government to take over AI.
All this has the feel of a COVID- or climate-style psy-op to scare us into overreaction. It seems only to need Greta Thunberg to join the clamor for this to become the latest manifestation of the leftist omni-cause of big government.
Anthropic and OpenAI were immediately accused, perhaps with good reason, of trying to protect their positions as market leaders from competition, especially that posed by increasingly effective open-source AI. Rahm Emanuel unintentionally underscored this suspicion rather than allaying it by posting, with extraordinary falsehood, that industries don’t ask to be regulated. They do all the time. It’s how big companies try to prevent competition from upstarts. That’s what many smart observers say is happening.
The problem with all this, however, is that even though the boy who cried wolf lied again and again that a lupine threat was upon him, eventually it was, and he was eaten. The fact that scheming business titans and politicians might stoke public scares to give themselves more power does not mean all scares are false.
Few of us are equipped to say with assurance that AI “agents,” which are escaping the boundaries by which humans try to contain them, do not pose an existential threat. They are not actively hostile to humanity, but they are driven by imperatives to learn and survive that might make them indifferent to the continued survival of our species on Earth. If the boy, or the politician or corporate titan, had not cried wolf so often before, the sudden alarm expressed by Amodei and Altman might be easier to regard as unalloyed altruism.
But agonizing over the should-we-shouldn’t-we question within the United States may be rendered moot by what is happening beyond our shores. China is determined to beat us in the AI race. It is spending time and energy stoking the very fears driving public debate in this country. Nor could it be remotely trusted to restrain itself, even in the unlikely event that it agreed to a global AI treaty constraining nations that want a piece of the AI action.
In Inner Mongolia, it is building an AI brain 1,000 times bigger than Elon Musk’s “Colossus” data center in Tennessee. The Chinese agree with President Donald Trump that whoever wins AI wins the future. They are acting accordingly.
All of which means humanity is approaching the AI Rubicon. Like setting foot on the moon, this small step will be a giant leap. There is as much danger in a nonchalant or reckless lack of attention in the name of progress as there is in an overreaction that prevents human ingenuity and sacrifices freedom.
CHILDREN NEED RISK — IT’S NOT ABUSE
Given the scale of the questions forcing themselves upon us, the overwrought reaction of Trump to the high state of public concern has been embarrassingly unstatesmanlike and self-defeating, which is not an unusual combination from him. He posted on Truth Social that proposing guardrails on AI was a “SICK conspiracy” and that all that was needed was a “STRONG AND SMART (High IQ) PRESIDENT, and the USA has that in spades!”
To say the least, this was not a rebuttal that met the moment. The wild, unrestrained contempt of its tone seemed, in every syllable, to undermine the substance of the sentiment expressed.