The AI Moral Panic Will Pass, Just Like All The Other Ones
From writing to the bicycle to the internet, history's great technology scares ended not in ruin but in human flourishing, and there is little reason to think AI will be the exception.
A few weeks ago, while chasing an unrelated story, we stumbled onto a quietly brilliant little website called Techlashed. It is an interactive timeline of technology-driven moral panics built by Andrew Maynard of Arizona State University’s Future of Being Human initiative, and he built much of it with AI, using the very technology now inspiring the loudest panic of our age to catalog every panic before it. Techlashed offers a much-needed deep historical perspective in the AI conversation
The pattern it lays bare is comically consistent, because every transformative technology arrives wrapped in dread, and that dread sounds the same every single time. Consider nine of them:
Writing. Yes, writing. In Plato’s Phaedrus, Socrates feared the written word would hollow out memory and leave us with the mere appearance of wisdom.
The printing press. Scholars warned cheap books would drown Europe in information and corrupt morals, when what followed was the Reformation, mass literacy, and modern science.
The bicycle. The 1890s warned of a disfiguring “bicycle face” and feared mobility would ruin our physical form.
The telegraph. Critics swore instant news would destroy leisure.
The Kodak camera. Eastman’s 1888 camera triggered a privacy panic over roving “Kodak fiends,” and yet photography did not end civilization.
Radio. A 1941 study found most surveyed children addicted to crime dramas, while moralists feared degenerate music was pouring into the family home.
Television. The same anxieties were rerun a generation later, certain the glowing box would dissolve the family altogether.
Video games. For decades the conventional wisdom held that consoles were breeding violent delinquents, a thesis the research has refused to confirm.
The internet and social media. Every prior fear bundled into one grand apocalypse, and yet here we all are.
Nine major panics, every one now so woven into ordinary life that the fear once attached to it reads as moderately to completely unhinged.
Which brings us to artificial intelligence and to the newest front in this very old war, the data center. The objections are familiar in shape, because we are told that AI will swallow every job, bury the culture in synthetic slop, and drink the rivers dry to feed its server farms. These worries are not nothing, but then neither were any of the others.
Take the data center anxiety on its own terms. Training large models does consume real electricity and water, and the build-out is straining some local grids, but the history of computing has been the story of doing exponentially more with exponentially less, and the room-sized machines of the 1950s burned far more power than the phone in your pocket that now outclasses them entirely. The firms with the most to lose from runaway energy bills are precisely the ones pouring fortunes into efficiency and new power, which is why efficiency here is not a pious hope but the most reliable trend the business has.
Then comes the oldest fear of all, that the machines will simply take our jobs. They may take some, because every general-purpose technology reshuffles the labor market, and that is genuinely painful for whoever is standing in the wrong place when it lands, but reshuffle is the operative word. The sewing machine was supposed to throw seamstresses into the gutter and instead built a garment industry that employed far more of them, while the automated teller machine was supposed to end the bank teller, yet teller numbers rose for decades as cheaper branches multiplied. The destruction is vivid and immediate while the creation is diffuse and slow, which is why the panic always reaches the masses first.
So could this time genuinely be different? Sure. But if we are placing bets, the historical ledger is lopsided to the point of being nearly one-sided. Each technology that was supposedly going to doom us instead quietly integrated into our lives and society moved on, and the fears were not so much crazy as badly aimed, fixed on the technology rather than on the temporary disruption it caused on its way to becoming ordinary.
A panic is not a signal that something has gone terribly wrong, but rather that the world is changing, and that we are once again doing what our species always does, which is to worry loudly, adapt quietly, and end up considerably better off than the worrying ever assumed. Artificial Intelligence may yet surprise us, and we should hold that with humility, but if you want to know how this kind of story tends to end, the past two thousand years have been about as consistent as history gets.
Well, many people, myself included, would argue that the last three items in your list are still systematically destroying our families and our culture. AI is a greater threat than all of these put together.
I say it often as someone in the construction industry involved in data center (DC) builds: The DC industry has itself to blame for much of this. Take water: Using open-loop evaporative cooling is incredibly wasteful. But it's cheaper, and it uses less power. So they get built. More advanced builds use chilled water. But many of those use cooling towers which, if open-loop, also use tremendous amounts of water. But they are electrically more efficient to operate than air-cooled, and therefore they get built.
There are easy engineering solutions to water and noise, but they are costly. Power is more difficult, but also workable, especially if DCs can be allowed to throttle back in times of peak need (less than 1% of the hours per year). Some operators are willing to do this, others - the same who want to use evaporative cooling - do not.
Eventually, the grid operators will stop allowing DCs to connect because they will run out of power slots. At that point, new builds will self-generate power, which they should be doing anyway. But again, this is costly. (and time consuming, and they want the new builds yesterday).
Focusing on long-term community partnership and solving water, power, and noise with tried and true engineering solutions - instead of squeezing every penny of profit with cheap builds - would go a long way towards helping the tech community overcome the panic.