Our Take
Goldman's comparison proves nothing about AI's labor impact, only that macro conditions improved—confusing causation with correlation and dodging the harder question of sectoral displacement.
Why it matters
As AI deployment accelerates through 2024, the labor market is the primary political and social vulnerability. Claims about overall health can mask concentrated job losses in white-collar work, making this narrative worth stress-testing now rather than after disruption hardens.
Do this week
Finance and HR leaders: audit your wage bill and headcount by role and tenure this quarter so you can quantify exposure to AI-driven role compression before it compounds.
Goldman Sachs reports labor market gains since ChatGPT's launch
Goldman Sachs economists found that the U.S. labor market has strengthened since November 2022, when OpenAI released ChatGPT. The analysis compares employment levels, unemployment rate, and wage growth in the months following the AI model's public debut against current conditions (per Fortune reporting of Goldman's research). The bank's conclusion contradicts widespread concern that generative AI would trigger rapid job displacement and wage suppression in the near term.
The claim is straightforward: if AI were causing measurable labor market harm, the aggregate data would show deterioration. Instead, Goldman found improvement across key indicators. This finding has circulated as a counterargument to AI-skeptics warning of imminent dislocation.
Aggregate data masks sectoral and skill-level concentration
Goldman's measure tells you almost nothing about where AI is actually displacing work. Aggregate employment and wage statistics smooth over regional shocks, occupational compression, and skill-level effects. If AI eliminates 500,000 junior analyst roles but creates 300,000 prompt-engineering and model-fine-tuning jobs at higher pay, the headline number looks fine. The workers displaced from junior roles do not.
The timing also matters. ChatGPT reached mass adoption in early 2023; corporate AI deployment at scale has only accelerated through late 2023 and 2024. Most labor market disruption from AI adoption has not yet flowed into unemployment or wage suppression at the aggregate level, but that does not mean it is not happening in specific sectors. Software engineering, legal research, customer support, and junior financial analyst roles are already seeing consolidation. The lag between deployment and labor market measurement is real.
Goldman's framing also assumes that if AI were harmful, the harm would show up uniformly across the economy. That is not how technological unemployment works. It concentrates first, shows as general deterioration later.
Treat aggregate health claims as incomplete
If you manage headcount, budget, or hiring in roles that involve routine analysis, research, or content production, Goldman's finding is a reason to pressure-test your own labor cost assumptions, not to relax them. The macro story covers the micro story. Audit your role-by-role exposure to AI substitution: which functions are most vulnerable to tooling that already exists or will exist in the next 18 months. Compare internal wage growth against external hiring velocity in those roles. If you are still hiring heavily into roles where AI can replicate 60% of the work, you are pre-loading a cost adjustment for later.