Unions win AI limits as 70% of US agencies use generative AI
US government unions win human oversight and protection from AI layoffs in many areas.
US government unions win more limits on AI at work in several states. An Ernst and Young survey says almost 70 percent of US government agencies use generative AI. This is software that writes text and makes content. In October, after a vote to stop work, a union in Oakland won training and not AI layoffs.
What unions won
- Oakland: IFPTE Local 21 won training and not AI layoffs.
- Pennsylvania: Governor Josh Shapiro promised SEIU Local 668 human oversight in a ChatGPT Enterprise pilot.
- Ohio: the Transport Workers Union won a veto on autonomous vehicles.
- California: Governor Gavin Newsom banned AI surveillance software and automated firings in September.
New York has protected public employees from AI displacement by law since late 2025.
The numbers
The one number is the Ernst and Young survey: almost 70 percent of US agencies use generative AI. The source gives no data on how the promises work in use.
What is not known
Labor researcher Kung Feng at UC Berkeley says employers push automation to cut costs. Liz Shuler of the AFL-CIO and other union leaders want AI safety systems that help workers. They want mandatory human control of each AI use at work.
Insight: Contracts, not laws, set the AI rules for staff
An October 2026 poll found that 66 percent of US voters want more AI rules, but lawmakers do not agree on a law. Unions win limits in deals, and the news union at USA Today wants a written promise on AI. This reads as unions that fill the space law leaves, and contract text can set the first rules for AI at work. You will see these terms at the table.
Sources
Posted