Adopted on 17 December 2025 and now published in the Official Journal, Parliament’s resolution formally invokes Article 225 TFEU to request a Commission legislative proposal on algorithmic management in the workplace, with a detailed eleven-recommendation annex setting out what it should contain
The resolution, adopted with reference 2025/2080(INL), marks the first formal legislative initiative by Parliament on AI in employment contexts. Under Article 225 TFEU, Parliament may request the Commission to submit a legislative proposal, and while the Commission is not obliged to comply, it must explain why if it does not. The resolution calls specifically for a Commission impact assessment complemented by a Competitiveness and SME test before any proposal is submitted, reflecting Parliament’s awareness that the file risks industry pushback if costs to SMEs are not demonstrably managed.
The factual context the resolution draws on is striking. Between one quarter and 80% of companies in the EU already use at least one form of algorithmic management, while 26.5% of workers report having their performance monitored by a computer programme and 27.4% have tasks allocated via automated systems. Only 18% of affected workers report receiving a detailed explanation of how those systems work or what their rights are, even though more than half of employers claim to have informed their staff. The resolution also flags that current generative AI investments have not yet produced measurable productivity gains in most corporate deployments, citing Gartner survey data.
The eleven recommendations in the annex constitute Parliament’s preferred legislative architecture. Recommendation 1 sets its scope as universal, covering every worker and employer in the Union, with equal protection for workers in intermediary contractual arrangements. Recommendation 3 on transparency requires employers to inform workers in writing, in accessible plain language, before their first working day, about any algorithmic management systems in use, covering purpose, the categories of data collected, human oversight mechanisms, and training measures available. Candidates in recruitment must receive equivalent information at the application stage.
Recommendation 5 lists prohibited practices. The prohibition covers: emotional, psychological or neurological state monitoring; surveillance of private communications; real-time geolocation tracking outside working hours; data collection predicting exercise of fundamental rights including the right to organise; and processing of special categories of personal data. Notably, the recommendation explicitly states that employee consent cannot provide a lawful basis for automated monitoring given the power imbalance inherent in the employment relationship, which would require an amendment to how Article 6(1)(a) GDPR operates in practice in this context.
Recommendation 6 on human oversight requires that all decisions on termination, renewal, remuneration changes, or disciplinary action must be taken by a human being and remain subject to human review, regardless of how AI-assisted the process is. Workers must have a right to obtain a meaningful explanation of any AI-influenced decision affecting essential aspects of their employment, and a right to request review. Recommendation 7 requires algorithmic management systems to be integrated into employers’ OSH risk assessment frameworks, covering psychosocial and ergonomic risks.
The resolution explicitly calls for coherence with the Platform Work Directive, whose algorithmic management provisions apply only to platform workers, by extending equivalent protections to all workers. Parliament also stresses that the proposal must not overlap or duplicate the AI Act, the DSA, or the GDPR, and must not create unnecessary administrative burden for SMEs (99% of EU businesses). The Commission must now respond to the resolution; a legislative proposal is not guaranteed but the political pressure is substantial.
