Poll Finds Strong Worker Support for Union-Backed AI Policies

More than nine out of 10 workers surveyed support policies on artificial intelligence that labor unions have sought, including requirements for human decision makers on issues affecting employment and guardrails against harmful uses of the technology.

The poll, conducted by David Binder Research for the AFL-CIO from 14 April to 22 April, found 95% of respondents back a rule that a human must serve as the final decision maker on any matter affecting individual workers and their jobs. Some 92% supported advanced guardrails against harmful applications of AI in workplaces along with transparency and accountability requirements for employers.

Every worker-protection measure in the survey received backing from at least 75% of those polled. This included 75% who favored expanding opportunities for workers to form unions to safeguard jobs from AI effects. The survey of 1,588 respondents across the United States also showed that 78% consider it extremely or very important for action to be taken to protect workers from potential harms of AI.

Only 7% said their employers have disclosed how and when AI monitors their work, while 70% reported no disclosure and 23% were unsure.

When asked which group is most trustworthy to protect workers from AI effects, 38% chose labor unions, 17% picked Democrats, 10% selected Republicans, 6% named employers and 18% chose none of the options. The AFL-CIO released the results as unions continue to negotiate AI-related protections through collective bargaining agreements.

One union member described companies turning to AI as a way to reduce staff and cut costs even when results are poor. Union contracts secured in 2024 have included provisions for editorial integrity, transparency on AI use and prohibitions on layoffs or pay cuts tied to the technology.

A registered nurse and union member in North Carolina said her contract requires union approval before any new technology affecting patient care can be implemented. The agreement also bars technology that would de-skill workers or undermine their professional judgment.

The nurse added that some hospitals have deployed AI for scheduling and for algorithms estimating whether patients will deteriorate, despite known errors in the models. She stated that patients should not be subjected to unconsented experiments with the technology.

“These results make it clear: our Workers First Initiative on AI is not just a set of principles, but a mandate to deliver." — Liz Shuler, president of the AFL-CIO (The Guardian) The polling comes amid growing efforts by workers to secure contractual limits on how artificial intelligence is deployed in workplaces.”