Workers need greater say over AI rollout, says TUC-backed report

Home › Workers need greater say over AI rollout, says TUC-backed report
Workers need greater say over AI rollout, says TUC-backed report

A May 2026 report from the Institute for Public Policy Research (IPPR), supported publicly by the Trades Union Congress (TUC), argues that workers should have meaningful influence over how AI is introduced at work. The report is a policy proposal, not enacted law, but its evidence is relevant to technology and governance teams.

Table of Contents

What the report found

IPPR reports that one in five surveyed workers said AI was making their job worse, one in 25 said they had already lost a job because of AI, and half said they had no say in workplace AI decisions. These are survey findings reported by IPPR; they should not be read as forecasts for every industry or country.

The report distinguishes between augmentation, degradation and displacement. AI can support workers, intensify or monitor work, or replace tasks and roles. Which outcome occurs depends not only on the technology but also on job design, incentives and who participates in deployment decisions.

What IPPR proposes

The proposals include stronger consultation rights, a worker-support levy and portable benefits that could fund representation, legal assistance or training. These are recommendations for UK policy. They are not current EU AI Act obligations and should not be presented as mandatory requirements for European employers.

What IT leaders can do now

  1. Consult before procurement. Ask affected employees and representatives which tasks cause friction, which data should be off limits and what failure would look like.
  2. Document the purpose. State whether the system is intended to assist, evaluate, monitor or automate. Avoid vague productivity claims.
  3. Assess people impacts. Review workload, surveillance, discrimination, accessibility, deskilling and appeal routes alongside security and financial risk.
  4. Keep consequential decisions reviewable. Assign accountable humans, define override authority and provide a channel for affected people to challenge outcomes.
  5. Measure both gains and costs. Track quality, rework, incidents, worker experience and distribution of benefits—not only time saved.

Unapproved tools can bypass both worker consultation and normal security review; see the practical guide to shadow-AI risk.

How this relates to European regulation

The EU AI Act is already in force and some provisions apply, including AI-literacy duties and Article 50 transparency obligations. Certain employment-related AI systems may be high-risk, but classification depends on intended purpose and the legal criteria. GDPR, employment law and national worker-participation rules may also apply independently.

A responsible rollout therefore combines legal review with genuine participation. Consultation is useful even where it is not specifically mandated: the people doing the work often identify poor data, unsafe shortcuts and unrealistic assumptions before they become production failures.

Sources

This article provides general information and is not legal or employment advice.

Editorial disclosure: AI tools may have assisted research, drafting or editing. ITnovati remains responsible for the published text. Time-sensitive technical, legal and product claims should be checked against the linked primary sources.