Study: AI Hasn’t Disrupted U.S. Jobs Since ChatGPT
AI hasn’t upended U.S. jobs. Not yet.
A new analysis from Yale University’s Budget Lab and The Brookings Institution finds the U.S. labour market has remained largely steady since the debut of ChatGPT, tempering predictions of an immediate AI-driven jobs crisis.
What the researchers found
The Yale report analyzing 33 months of U.S. labour data since OpenAI’s release of ChatGPT in November 2022 found that fears of widespread AI-driven unemployment remain “largely speculative.” Researchers studied employment trends across workers with high, medium, and low exposure to AI and observed little to no shift in overall workforce composition.
The study did note a decline among early‑career workers, but the authors cautioned that broader economic dynamics—not AI alone—could explain the change.
Why AI’s impact appears muted
Workplace adoption of AI varies widely by sector, with practical barriers slowing deployment:
- Privacy and security risks
- Data governance challenges
- Need for high‑quality, well-labelled data
- Organizational change management and training
- Uncertain regulatory and compliance requirements
Context and contrasting forecasts
The findings diverge from some headline projections. The World Economic Forum’s 2025 Future of Jobs Report projects 92 million roles could be displaced by AI, and Anthropic CEO Dario Amodei recently warned that more than half of entry‑level white‑collar positions could be automated.
By contrast, the Yale–Brookings team argues that transformative technologies typically take years—often decades—to fully reshape work, requiring complementary investments, cultural change, and regulatory clarity.
What this means for policymakers and employers
Their message is not that AI won’t change jobs, but that the immediate impact appears muted. The authors urge continuous monitoring of labour trends and greater transparency from major AI developers to help businesses and policymakers discern real risks from hype.
Without robust evidence, they warn, companies could overreact to perceived threats or miss emerging disruptions.
Bottom line: AI’s long‑term effects on employment may be significant, but early data point to gradual change rather than a sudden shock.
The road ahead
Policymakers, researchers, and employers will need sustained, evidence‑based assessments to navigate what could be one of the most consequential technological transitions of this era.
