aceUS workers overwhelmingly support union-backed policies on AI, poll says
US workers overwhelmingly support pro-worker policies on artificial intelligence (AI) and view labor unions as the most reliable protectors of workers from the effects of AI, according to a new poll released by the AFL-CIO, the largest federation of labor unions in the US. More than nine out of 10 workers surveyed expressed support for policies on artificial intelligence that labor unions may fight for, including 95% supporting a requirement that a human be the final decision maker on any issues affecting individua
US workers would gain workplace protection from AI through union-backed policies requiring a human to remain the final decision maker on issues affecting individuals, with broad survey support for such rules.
Evidence
- Peer-reviewedJournal of Responsible Technology2020-10-16
- Peer-reviewedEuropean Journal of Risk Regulation2025-10-29
- Peer-reviewedSafety Science2023-09-23
- JournalismThe Guardian2026-05-12
- Peer-reviewedInternational Journal of Human–Computer Interaction2023-01-02
How should this claim be treated?
Truvace Impact Record TRV-2026-0091, v9: “US workers overwhelmingly support union-backed policies on AI, poll says.” Truvace, 2026-10-10. /record/TRV-2026-0091 (accessed at citation time). sha256 a5a6e36b38dbf78a…
Calibration history
Every change to this record since certification, in the open.
Source set updated
Source set updated
Source set updated
Source set updated
Model backfill: grounded claim, summary, sector, and trace validation
Model backfill: grounded claim, summary, sector, and trace validation
Restored after model confidence scale normalization
Model backfill: source did not support a publishable AI-impact claim
Certified into the record
How to verify without trusting this page
Fetch the canonical text of any version from /api/record/TRV-2026-0091 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.