The political agreement reached yesterday by the co-legislators on the AI Omnibus marks a dangerous turning point for European industrial workers operating machinery. IndustriAll Europe represents workers in the mining, manufacturing and energy sectors, millions of whom work in increasingly automated environments, where AI technologies are becoming ever more embedded in machinery, automated systems and personal protective equipment (PPE).

For industriall Europe, the regrettable message behind this deal is clear: when industrial lobbying intensifies, workers’ rights and occupational safety protections become negotiable.

From the beginning of the discussions, we warned about the serious consequences this deregulation agenda could have for workers and for occupational safety and health (OSH), particularly through the sectoral legislation listed in the AI Act, such as the Machinery Regulation and PPE rules, as well as through the weakening of the definition of “safety components”. Despite these warnings, the final compromise confirms many of the concerns raised by trade unions and worker representatives.

One of the most alarming elements of the agreement is the decision to move the Machinery Regulation from Annex I, Section A of the AI Act (which ensures the strongest and most harmonised safety requirements) to Section B instead. This is not a technical adjustment. In practice, Section A guarantees that high-risk products automatically fall under the strictest AI safety obligations across the EU. Moving machinery into Section B means these safeguards are no longer applied in the same automatic and consistent way. As a result, many AI systems embedded in industrial machinery will no longer be systematically covered by the strongest level of protection under the AI Act, leading to a weaker and more fragmented safety framework for workers across Europe.

This creates a major regulatory gap for industrial AI systems. Even with the so-called safeguards presented by negotiators, the outcome establishes a dangerous precedent. The result is greater legal uncertainty, fragmented enforcement, and an uneven playing field across Europe. It also adds further legal complexity, running directly counter to the objective of “simplification”.

This is part of a broader and already well-established trend. Already during the 2023 negotiations on the Machinery Regulation, the explicit reference to “high-risk machinery” was removed, reflecting the influence of industry lobbying in favour of lighter obligations and fewer controls. In the same process, the number of machinery categories subject to third-party conformity assessments was progressively reduced from 25 to just six in the final regulation.

Responsibility for this AI Omnibus outcome and the general deregulation process is shared across the political spectrum. The European Commission, some Member States and certain political groups in the European Parliament all contributed to advancing a narrative that frames regulation as a burden rather than as a democratic guarantee of safety, accountability and workers’ rights. The implications go beyond AI policy alone. At stake is not only regulatory coherence, but also the credibility of the European social model itself.

Isabelle Barthès, industriAll Europe General Secretary, said: "Workers in mining, manufacturing and energy sectors already face hazardous and increasingly automated working environments. Weakening AI safeguards at this moment sends the wrong political signal and undermines confidence in Europe’s commitment to safe and human-centred technological transition."

IndustriAll Europe will continue to closely monitor developments, including in the upcoming standardisation and comitology processes, where many decisive implementation choices will still be made.