These days, the conversation around AI automation and the job market is increasingly focused on “labor displacement,” the phenomenon in which new technology eliminates certain jobs but supposedly creates new ones elsewhere.

But AI, more than any tech that came before it, represents the possibility of mass unemployment on an unprecedented scale. Since workers in market economies depend entirely on employment for survival, mass unemployment would leave untold millions of people without anything to lose. Whether AI actually causes that remains a topic of debate, but the outcome if it does could be widespread social upheaval.

working class person’s wellbeing in a market economy depends almost entirely on their employment, it’s entirely possible that — if mass AI unemployment ever does genuinely hit, which to be fair is a major “if” — the conditions for serious social unrest could emerge.

In the hotbed of dog-eat-dog capitalism that is the United States, anti AI-sentiment is already on the rise. The data centers underlining the AI boom are widely reviled, and a surprising number of workers are admitting to sabotaging their company’s AI in the workplace. According to one survey, seven in ten people living in the US already think AI will make it harder to find work, a sentiment that isn’t helped much by a horrible job market.

As Royal Military College of Canada political scientist Yannick Veilleux-Lepage argued in a recent paper on AI and populist backlash, “AI generates the structural conditions historically associated with the onset of political violence.”

That discontent, he notes, stems from increasingly undemocratic decisions: data centers forced on small towns without consent, nonstop surveillance by corporate security firms, and major government handouts for tech industry projects, to name just a few.

“As AI company executives acquire more personal security, risk may shift to researchers on open campuses; as corporate campuses harden, risk shifts to the power substations that serve them; where national figures are unreachable, local policymakers who approved the data center become the proxies for the same structural anger,” Veilleux-Lepage writes.

As the Atlantic notes in a recent piece discussing AI backlash, tech executives are already walking back their apocalyptic rhetoric. Where the AI industry once bragged about the coming AI job apocalypse to boost their valuations, they’re now downplaying the risks of automation as grievance grows.

In 2023, for example, the Atlantic notes that OpenAI CEO Sam Altman bragged that “jobs are definitely going away, full stop.” These days, his prediction is much more even-keeled: “jobs doomerism is likely long-term wrong,” Altman wrote on social media.

Perhaps he’s covering his bases after his San Francisco estate was firebombed by an angry 20-year-old, or maybe his beliefs have really changed. Either way, his problems aren’t going away: it isn’t about what executives say, but about what they do.

More on AI and labor: Large Study Finds That Replacing Workers With AI Is Backfiring Badly