Latest / Minds, Bodies, and Terawatts / Goldman Sachs Warns of Lasting 'Scarring' Wage Cuts from AI Displacement

Transcript

Goldman Sachs Warns of Lasting 'Scarring' Wage Cuts from AI Displacement · Minds, Bodies, and Terawatts

• 0:00
  [Host] This is Minds, Bodies, and Terawatts, where Wall Street quantifies what your paycheck already told you. Business Insider reported on April 7th that Goldman Sachs, in their report "AI's Displacement Dilemma: Labor Market Scarring in the Age of Automation," analyzed 40 years of labor market data and concluded that AI-driven displacement could impose lasting earnings losses on affected workers. Permanent wage scars. But there's a twist buried in the data that flips the generational anxiety narrative.
• 0:28
  [Guest] Goldman's research team went back four decades through displacement events, recessions, industry collapses, trade shocks, and mapped what happens to workers' earnings after they lose jobs to structural shifts. The finding that matters: displaced workers don't just take a temporary hit. They earn less for years, sometimes permanently.
• 0:45
  [Host] So we're talking about someone who loses a $75,000 job and lands a $58,000 job, and five years later they're still at $60,000 instead of climbing back?
• 0:56
  [Guest] That's the pattern Goldman found across four decades. The term they're using is "scarring," borrowed from labor economics literature on recessions, but applied here specifically to technology-driven displacement. When your skills become less valuable because a machine does them cheaper, the market reprices you permanently.
• 1:12
  [Host] And this isn't hypothetical. Challenger's latest data shows 2025 layoffs running 58 percent above 2024 levels, with January 2026 hitting 108,000 announced cuts, worst January since 2009.
• 1:22
  [Guest] Goldman's contribution is putting historical weight behind what's unfolding right now. We've had anecdotal evidence, but 40 years of displacement data gives you statistical confidence. The scarring effect held across industries, across decades, across different types of structural change.
• 1:37
  [Host] Goldman also found that younger workers may be less at risk than feared. That surprised me.
• 1:43
  [Guest] It's counterintuitive until you look at the mechanism. A 52-year-old financial analyst who's spent 20 years building expertise in a particular modeling approach faces a steeper cliff when that approach gets automated. Their wage premium was built on that specific expertise.
• 1:58
  [Host] Because younger workers haven't accumulated that kind of specialized capital yet?
• 2:02
  [Guest] Exactly. They have less to lose and more flexibility to redirect. The scar heals faster when you're 24 than when you're 52 with two decades of narrowly specialized human capital.
• 2:13
  [Host] But the Dallas Fed's working paper, "Artificial Intelligence and the Decline of Entry-Level Employment," found workers aged 22 to 25 in AI-exposed occupations saw a 13 percent employment decline since 2022. Software developers in that age group dropped nearly 20 percent from their late-2022 peak.
• 2:30
  [Guest] Employment access and wage scarring are different problems. Young workers are struggling to get hired. Entry-level postings dropped 16 percent on Handshake while applications per opening jumped 26 percent. The pipeline is clogged.
• 2:42
  [Host] So the barrier is getting in the door, not recovering once you're through it?
• 2:47
  [Guest] Right. A 24-year-old who gets laid off and pivots has a longer runway and lower switching costs than a 50-year-old with a mortgage and kids in college. The initial wound is real, but Goldman's data says the trajectory afterward diverges sharply by age.
• 3:03
  [Host] So the good news for young workers is "you'll recover from getting fired more easily." Cold comfort.
• 3:09
  [Guest] And Axios added an important caveat. Morgan Stanley's analysis, "Separating AI Signal from Corporate Noise," cautions that executives may be overstating AI-driven cost-cutting to impress investors. Some layoffs attributed to AI efficiency may be conventional cost reduction dressed in AI language.
• 3:25
  [Host] Because the market rewards the AI narrative?
• 3:27
  [Guest] If you announce layoffs and say "restructuring," your stock yawns. If you say "AI-driven efficiency transformation," your stock pops. Challenger data showed AI ranked only sixth as a stated reason for 2025 layoffs, well behind "restructuring" and "economic conditions."
• 3:42
  [Host] That gap between stated reasons and actual reasons makes Goldman's scarring data harder to interpret.
• 3:47
  [Guest] It does. How much displacement is genuinely AI-driven versus conventional cost-cutting wearing an AI costume? And that muddies the policy response. If you're building retraining programs for AI displacement but half the layoffs are old-fashioned downsizing, you're aiming at the wrong target.
• 4:03
  [Host] The workers don't care about the taxonomy though.
• 4:06
  [Guest] Whether you lost your job to a language model or to a CFO chasing a better quarterly number, the mortgage payment is the same. The scarring is the same.
• 4:17
  [Host] What does this mean for the 89 percent of 2026 graduates who, according to Intelligent.com's survey, fear AI will eliminate their jobs before they even land one? That's the anxiety the unscarcity dot ai article "Gen Z: 89 Percent of 2026 Grads Fear AI Will Take Their Jobs" digs into.
• 4:35
  [Guest] The survey data is striking. Over half of Gen Z already hold side gigs, according to Bankrate. Only 40 percent believe a four-year degree is worth the cost, per Gallup. They're building capability portfolios instead of collecting credentials, betting on things AI can't replicate: complex negotiation, care that feels human, ethical judgment in ambiguous situations.
• 4:55
  [Host] But there's a ceiling on individual adaptation. If the structural economy is repricing human cognitive labor downward, individual hustle doesn't change the macro trajectory.
• 5:04
  [Guest] That's Goldman's implication, even if they didn't say it explicitly. Scarring is a systemic phenomenon. It means the labor market itself is broken for displaced workers, not that they individually failed to adapt.
• 5:17
  [Host] Goldman's 40-year dataset raises a question nobody's answering well. Previous displacement events eventually created new sectors that absorbed workers. Is there evidence AI does the same at comparable wages?
• 5:27
  [Guest] Not yet. And Goldman's historical data argues against easy optimism, because the scarring effect held even during periods when new industries were emerging. Workers displaced from manufacturing in the 1990s didn't fully recover their earnings even as the tech boom created millions of jobs. Those new jobs went to different people with different skills in different places.
• 5:49
  [Host] So the macro economy grows while individual workers permanently lose ground.
• 5:53
  [Guest] Both things are true simultaneously. The economy gets richer. The displaced get poorer. That divergence is what the scarring data actually measures.
• 6:01
  [Host] Patrick Déglon calls that exact divergence the core tension in Unscarcity, the gap between aggregate abundance and individual precarity. For more on wage scarring, generational displacement, and the skills that stay human, head to unscarcity dot ai. The book is on Amazon, the audiobook is on Spotify. This has been Minds, Bodies, and Terawatts.