rade unions are raising serious concerns that the provisional agreement reached on the AI Digital Omnibus risks weakening protections for workers and undermining the EU’s safety framework for artificial intelligence.

The ETUC is particularly concerned by the removal of the Machinery Regulation from the scope of the AI Act. In practice, this would exempt AI-powered safety components that are embedded in industrial applications from complying with the essential requirements of the AI Act.

This proposal is unnecessary, as the AI Act already incorporates mechanisms designed to prevent administrative duplication. Rather than simplifying compliance, the proposed exclusion fragments the EU’s safety architecture and effectively hollows out a key pillar of the EU’s horizontal approach to trustworthy and safe AI.

The ETUC also warns against the further narrowing of the definition of a “safety component”. Under the AI Act, only AI systems intended to function as safety components and embedded in products subject to third-party conformity assessment qualify as high-risk.

Further limiting this scope, creates a substantial risk that AI systems embedded in products used by workers could be placed on the market or put into use without being classified as high-risk. As a result, these systems risk bypassing minimum safeguards, leaving workers exposed and unprotected.

Isabelle Schömann, ETUC Deputy General Secretary said:

“Employment remains a highly sensitive domain for the deployment of artificial intelligence. There are good reasons why the workplace is identified as high risk in the AI Act and requires robust safeguards.

“As AI systems are increasingly integrated into managerial functions that directly and may adversely affect workers throughout the entire employment relationship. Worker’s health and safety cannot be compromised.

“Key protections originally established in the AI Act, specifically those pertaining to the employment context need to be preserved and cannot be thrown under the ‘simplification’ bulldozer.”