The case for legislation dedicated to how artificial intelligence is used at work was strengthened today by a European Parliament vote which creates further uncertainty on this crucial issue.

In 2024, the EU adopted its first legislation on artificial intelligence (the AI Act) which recognised the high-risk nature of the use of AI in the workplace. However, the law is now being amended through the ‘AI Digital Omnibus’.

The report adopted by the Parliament today upholds key elements of the original act, including the requirement for providers and deployers to promote the AI literacy of their staff. However, the report also introduces amendments that weaken parts of the original Act which are of particular relevance to workers’ safety.

The European Trade Union Confederation therefore calls for the forthcoming Quality Jobs Act to include legislation on the use of AI at work which will ensure:

• The human in command principle, including the right for workers to challenge and overturn automated decisions;
• Transparency and explainability, including clear information on the use, logic and impacts of AI systems, and access to relevant data for trade unions and workers’ representatives;
• Stronger collective rights, with mandatory involvement of trade unions and enforcement of information, consultation, participation and training rights;
• A ban on intrusive practices, such as the processing of non-work-related personal or psychological data;
• Protection of health and safety, with mandatory risk assessments before deployment and accountability for harm caused by AI systems;
• Measures to prevent market concentration, including the application of competition rules and support for European technological sovereignty.

ETUC Deputy General Secretary Isabelle Schömann said:

“The AI Act was an important first step towards putting responsible limits on the use of artificial intelligence. However, it did little to ensure that the rapid rollout of AI in workplaces is done in a way that respects workers’ rights and upholds safety standards.

“Some of the already limited references to the workplace in the AI Act are now at risk of being weakened through the Digital Omnibus. That has made it even clearer that we need legislation dedicated specifically to the use of artificial intelligence in the workplace through the Quality Jobs Act.

“Artificial intelligence is already changing the world of work. It has the potential to make workplaces more productive and safer – but only if EU policymakers ensure that is the case. The Commission’s own data shows huge public support for rules needed to protect workers from the risks posed by AI. What’s more, research shows that workers are more likely to report positive impacts of AI if they have been consulted on its use through their trade unions.”