Automation
New York AI Job Risk Model Shows 628,000 Positions Exposed
Cornell Tech and Appleseed built an interactive calculator revealing AI's employment impact falls between mild recession and pandemic-level disruption.
Key takeaways
- Cornell Tech and Appleseed's AI Job Exposure Calculator models 628,000 New York jobs at risk under most likely scenarios—between recession and pandemic impact levels.
- Only 4% of New York service firms using AI report layoffs, despite 61% adoption, revealing a gap between automation capability and actual displacement.
- Monte Carlo simulation with 20,000 iterations shows worst-case scenarios require unlikely simultaneous maximums in adoption, displacement, and minimal worker protection.
- Recommended policies prioritize worker mobility over job protection, unemployment insurance strengthening, and voluntary wage standards to share AI productivity gains.
- The model calculates exposure magnitude rather than timing, leaving uncertainty about whether risks materialize in months or years.
A simulation tool cuts through AI employment hype
While 61% of service firms in the New York region now use AI—up from 25% in 2024, according to September data from the New York Fed—only 4% report layoffs tied to the technology. Yet major employers from Microsoft to Kroger continue announcing AI-related cuts, leaving policymakers caught between contradictory signals about automation's real impact.
To ground the debate in evidence rather than speculation, Cornell Tech's Urban Tech Hub and economic development firm Appleseed created the AI Job Exposure Calculator, an interactive tool that models how different AI adoption scenarios could affect New York State's workforce. The project emerged from work on Lower Manhattan's economic future and supports Governor Hochul's FutureWorks commission, which is tasked with preparing the state's workers for AI disruption.
The calculator processes current employment across 845 occupations in New York's 10 labor markets, then applies variables for AI exposure (drawing on Yale Budget Lab's synthesis of six measurement methods), actual employer adoption rates, displacement percentages, and protective effects where AI helps workers stay productive rather than replaced.
Why it matters
Without tools to model realistic scenarios, policymakers risk either under-preparing for genuine displacement or wasting resources on overblown fears. The calculator reveals that while AI job risk is real and substantial, the apocalyptic forecasts dominating headlines require multiple worst-case assumptions to materialize simultaneously—an unlikely convergence.
Monte Carlo analysis reveals most probable outcome
The research team used Claude Code to build a Monte Carlo simulation that runs 20,000 iterations with randomized variables. The most likely middle-ground scenario—57% adoption, 30% displacement, 52% protection—threatens approximately 628,000 jobs statewide, or 6.3% of the labor force.
That figure sits between New York's two biggest recent shocks: the 2008-2009 recession eliminated roughly 335,000 jobs over 18 months, while COVID-19 erased nearly 2 million positions in weeks during March-April 2020. The distribution curve shows extreme outcomes require simultaneous maximum adoption, displacement, and minimal protection—a statistically improbable alignment.
Importantly, the model calculates exposure size rather than timing. The 628,000 figure represents jobs at risk if given adoption levels occur, whether in three months or three years.
Three policy priorities for New York
The researchers recommend focusing on worker mobility rather than job scarcity, noting that long-term growth could offset losses even as millions need occupation changes. This requires streamlined state licensing, reduced business formation barriers, portable benefits, and expanded AI training through employers, colleges, libraries, and programs like the state's AI Jobs Institute targeting 40,000 workers.
New York should also strengthen unemployment insurance beyond recent improvements—repaying pandemic loans, raising taxable wage ceilings, and increasing maximum weekly benefits to $869. AI displacement combined with demographic shifts and trade disruptions could strain the system within years.
Finally, the state must ensure workers capture AI productivity gains. The analysis points to voluntary minimum wage standards in finance and other exposed sectors, citing how Bank of America, JPMorgan Chase, and others raised hourly floors to $20-25. Research shows these standards lift wages for covered workers, those slightly above them, and competitor employees. The report also suggests extending equity compensation enjoyed by AI company employees to contract workers in maintenance, security, and food service.
These recommendations and the underlying calculator methodology are detailed in analysis first reported by Vital City NYC, which notes that AI's workforce effects are evolving too rapidly for static forecasts—requiring continuous monitoring and adaptive policy responses.