Public-sector unions in New York supported a law, effective December 2025, that protects workers at state agencies from being displaced by AI. Public-sector workers have also backed AI-related legislation in other states.
The union voted on 22 October to authorize a strike, with 94.1% in favor, after pushing for protections to ensure members would not be laid off or disciplined by AI.
The Bay Area union is one of many public-sector labor groups nationwide seeking explicit protections against AI and automation. Union representatives say the tools could worsen working conditions and threaten jobs.
Unions supported another California bill that prohibited employers from using workplace AI surveillance tools, including systems that could monitor employees’ emotional states.
Researchers call some of those systems bossware: AI tools that public- and private-sector employers are increasingly rolling out to monitor and supervise employees, according to an October report co-authored by Feng. Some tools rank and evaluate workers for hiring, firing and promotions.
Workers’ requests include advance notice before a new technology is rolled out, information about that technology, and the right to negotiate over its use.
Some automation could aid public-sector workers, including vehicle safety systems in public transit, but humans should always be in charge and provide oversight, Gonzalez said. In transport, union representatives said the technology is already in front of them and that they are fighting it every day.
Employers in the public sector often deploy automation during claimed attempts to improve efficiency and cut costs when budgets are slim, Feng said. Unions have faced mass job loss before, including when automation on assembly lines in the Midwest reduced manufacturing jobs, a union representative said.
One state encouraged employees, including unemployment-claims workers and social workers, to use ChatGPT Enterprise to test for time savings. Critics of rapid adoption have argued that a new system can cost more, be harder to implement and produce weaker results.
Unions have been pushing for AI protections since the late 2010s, and those battles are intensifying as automation advances and generative AI becomes more common, according to one account of the dispute. Concern has grown since ChatGPT’s debut in 2022.
“It’s powerful for elected leaders to say, there will always be human public servants under my watch,” Catanese said, referring to commitments that AI will not displace workers. Union representatives have said they are using contract negotiations to institute guardrails that policymakers will not.