AI Expert Says It’s Time to Stop Freaking Out About AI Taking Our Jobs

A majority of US voters surveyed believe AI will increase unemployment, but NYU cognitive scientist Gary Marcus argues AI-driven mass job loss is unlikely in the near term. Company data cited: Anthropic found "no systematic increase in unemployment for highly exposed workers since late 2022," and Klarna reversed its automation-driven layoffs 11 months after announcing them. Marcus contends firms may be "AI-washing" investor communications to boost valuations or obscure performance-driven layoffs.

Analysis

Markets are mispricing the difference between infrastructure-driven productivity and durable headcount elimination. Expect a multi-step adoption curve: large enterprises will spend on compute, tooling and pilots over the next 6–24 months (supporting vendors), but visible permanent net layoffs from AI are unlikely to be a broad, sustained phenomenon within a 12-month window; I estimate <25% of public “AI-enabled” headcount reductions will remain permanent past 12 months as firms backfill skills or reverse decisions once customer / operational friction appears.

That creates clear second-order flows: AI infrastructure (GPU/TPU providers, cloud) will capture capex regardless of near-term employment outcomes, while firms that used “AI” as a rationale for cost-cutting will suffer governance, customer-service and regulatory repricing. Expect above-average volatility in exposed fintechs and customer-service heavy platforms for 3–9 months as credibility and regulatory scrutiny are re-priced; a 20–40% valuation swing is plausible for smaller-cap names with recent automation narratives.

Tail risks are asymmetric: a credible, measurable automation product that demonstrates >15% sustainable FTE substitution across multiple enterprise customers would force a rapid rerating within 12–24 months, but that scenario remains low-probability in the near term. Key catalysts to watch are (1) audited productivity metrics from enterprise pilots over the next 6–12 months, (2) regulatory guidance or enforcement on AI-driven workforce changes within 6–18 months, and (3) quarterly capex commentary from hyperscalers which will determine 12–36 month upside for infrastructure names.

Key Decisions for Investors

• Short KLAR (KLAR) 3–9 month horizon: size to 1–2% NAV. Rationale: governance/reputational risk and operating reversals likely compress multiples; target 30% downside, stop-loss at +15% to control headline-driven rallies.
• Buy NVDA Jan-2027 LEAP calls (NVDA) as a core 12–36 month convexity trade: infrastructure demand should outlast the hype cycle. Risk/reward ~ asymmetric: pay known premium for optionality, target >2x if data-center spend continues; hedge by selling shorter-dated calls if implied vola spikes.
• Long ManpowerGroup (MAN) 6–12 months: tactical long to capture rehiring, outsourcing and retraining demand as firms reverse premature automation. Expect 15–30% upside if labor markets normalize; defensively hedge with a small put position for event risk.
• Pair trade: long MSFT (12–24 months) / short KLAR (3–9 months) to isolate secular cloud/AI infra exposure versus idiosyncratic fintech governance risk. Weight to neutral market-beta; target pair P&L positive if cloud guidance holds while fintech credibility is re-priced.