The Techlash is Silly. We Still Have to Address It.
Predictions that the sky is falling are wrong; they're also making it hard to push forward on important new technology.
In politics, it’s always hard to predict what will make people mad next. Turn the clocks back a few years and almost no one would have expected that voters would be outraged—literally flocking to town halls—over the construction of data centers for training and deploying artificial intelligence. Even Sen. Bernie Sanders has gotten in on the action, calling for a national moratorium on data-center construction and cheering cities like Denver that have already implemented one.
It’s not just data centers. New York Gov. Kathy Hochul recently killed her own budget proposal to allow driverless taxis outside of New York City after she faced dissent in Albany. Janeese Lewis George, a leading candidate for mayor of Washington, D.C., has also come out against autonomous cabs. And voters across the political spectrum are increasingly aggressive in their support for regulating AI of all stripes.
At the level of policy, this “techlash” is a mistake. Not all technological innovation is good, but our leaders are underestimating the benefits, and overestimating the costs, of the current crop they’re opposing. At the same time, though, anti-tech politics are a predictable consequence of how some of the most eager advocates have presented these tools. A rhetorical rethink is seriously needed.
Today’s techlash politics are mostly driven by a combination of misplaced fears and entrenched interests taking advantage of those fears. Take the deployment of autonomous vehicles. Car accidents are a leading cause of non-medical death, claiming over 40,000 lives per year. Many people are afraid that the absence of a human driver will increase this risk, explaining why two-thirds of respondents to one poll said they were “afraid” of driverless vehicles.
In reality, autonomous vehicles make our roads significantly safer. Research has found that Waymos—an increasingly popular autonomous-vehicle taxi service—are as much as 96 percent less likely to be in injury-involved crashes than are similar human-driven vehicles. Driverless vehicles also free up commuter time, potentially saving workers hundreds of hours a year. This is perhaps why support for autonomous vehicles has surged in San Francisco as use has grown.
Nonetheless, fear is a potent political force. Hochul’s about-face on driverless vehicles was cheered by an Uber-supported drivers’ group, Streetsblog notes. Such incumbents are often a key mover behind techlash politics.
What about data centers? Many are concerned about their effects on the environment and consumer costs. As Sanders put it, “data centers will have a profound impact on land and water use, and will drive up electricity costs.”
As the writer Andy Masley has demonstrated, concerns about water use are “fake”: U.S. data centers use something like 0.04 percent of America’s fresh water, equivalent to just 6 percent of the use of the golf industry. Concerns about energy are slightly more legitimate: data centers used 4 percent of America’s energy last year, Pew estimates, with that figure likely to rise precipitously in coming years. But, as Craig Roach and John Garnett recently argued in City Journal, the solution to rising energy demand is deregulation, not moratoria and hand-outs.
Moreover, there are real economic upsides to data centers. Constructing them adds jobs in the short run, although over the long run the effects appear to be negligible. States that work to attract them—like Texas and Virginia—can conceivably extract bigger co-investments from rapidly growing AI firms, expanding these benefits.
Where Sanders and other techlash proponents have a more potent case is in their bigger fears about the ways technology could radically reshape the workforce. As Sanders put it, “AI will likely have a catastrophic impact on the lives of working-class Americans, eliminating tens of millions of blue- and white-collar jobs in every sector of our economy.” These are concerns animating opposition to driverless cars, too—what will they do to the hundreds of thousands of jobs created by Lyft, Uber, and taxis?
Which is where the most vocal advocates of these new technologies have been shooting themselves in the foot. As pollster Nate Silver recently observed on X, many voters are responding not just to their irrational fears, but to what the actual people inventing these new technologies are telling them.
Anthropic CEO Dario Amodei, for example, predicted last year that AI will soon take half of all white collar jobs. A recent viral X post by AI investor Matt Shumer analogized the current moment to February 2020, with AI on the verge of being as disruptive as the COVID-19 pandemic. When Sanders says that he expects AI to have a “catastrophic impact,” he’s just repeating what some of the people building AI are saying.
He might have a point, too, if AI were likely to topple the U.S. economy. But it probably isn’t: a recent survey of 6,000 business leaders found that they project AI will cause a small decline in employment—0.7 percent—over the next three years, in combination with a 1.4 percent increase in productivity.
Even those numbers may be too pessimistic: automation often increases human employment as new productivity creates new job opportunities. That’s why, for example, the spread of the Automated Teller Machine (ATM) led to an increase in the number of human bank tellers, as lowered costs allowed for the opening of more bank branches and thus more employment.
There are, of course, serious arguments against this rosy view. Technological innovation is not everywhere and always good, and policymakers should be making prudential judgements about the costs and benefits of new tech, in terms of both its economic impacts and the ways it can change the human condition.
But the techlash does not seem to be powered by such prudential judgements. Rather, it’s in large part a product of the chicken-little fears that many advocates of new tech are themselves pushing at every available opportunity.
The political consequences of such rhetoric are now clear: fewer data centers, more road deaths, less growth, less human well-being. It would be a shame if we kept going down this road—as it seems an increasing number of politicians are looking to do. If those who stand to benefit the most from new tech want to avoid that outcome, they need to turn down the alarmism, and focus instead on all the ways these products can make everyone better off.
Data centers are left in an obvious double peril
Citizens do not want data centers that increase consumer power prices
But many citizens are also opposed to permitting new power generation
We've made a lot of progress on culturally convincing people to oppose new housing construction less strongly and less often but that hasn't really spilled over into less opposition for industrial building