Algorithmic discrimination and ranking: collective protection and compensation for trade unions

1. The fact

   A first-instance decision (Trib. Bologna ord. 31.12.2020) ascertained the discriminatory nature of the ranking system used by a well-known food delivery multinational. The facts stem from the use of shift management software that assigned reliability scores based on the worker's constant availability. According to the claims supported by trade unions, the algorithm penalized riders who were absent due to illness or strikes, thereby reducing their access to future, more lucrative shifts. The proceeding saw the recognition of liability for indirect discrimination, with a specific focus on the compensation for damages awarded to the applicant trade unions, quantified at 50,000 euros.

2. 

   The rules at play

   The regulatory framework hinges on several sources:

   1. Legislative Decree 198/2006 (Equal Opportunities Code), Art. 25, which defines indirect discrimination as any apparently neutral provision or practice that puts persons of a protected group at a disadvantage, and Art. 40 on the reversal of the burden of proof.
   2. Legislative Decree 150/2011, Art. 28, which regulates disputes concerning discrimination.
   3. EU Regulation 2016/679 (GDPR), Art. 22, which prohibits decisions based solely on automated processing that significantly affect the individual without adequate safeguards.
   4. The AI Act (EU Regulation 2024/1689), which classifies worker management systems as high-risk, imposing human oversight and transparency obligations regarding input data.

3. What the case law says

   Courts of merit have clarified that the neutrality of an algorithm is a technical illusion where its effects result in a violation of constitutional rights. The ruling of the Court of Bologna dictates that the employer cannot hide behind software complexity or trade secrets to deny access to ranking logic. In particular, judges have ruled that once the worker or the trade union provides statistical evidence suggesting discrimination, the burden of proof shifts: the platform must prove that the parameters are non-discriminatory. Regarding compensation, case law has recognized damages in favor of trade unions for the infringement of trade union prerogatives and the right to strike.

5. Analysis drafted and verified with edit.legal

   To verify the provisions cited in this article, we used edit.legal. Test our legal AI on official sources and apply it to your own matters.

   Try edit.legal AI

8. What it teaches professionals

   1. Adopt an evidentiary strategy based on the analysis of statistical data provided by the worker in order to trigger the reversal of the burden of proof.
   2. Challenge the lawfulness of ranking systems that lack human intervention for exceptions connected to fundamental rights.
   3. Consider collective trade union action to claim compensation for damages resulting from the infringement of trade union prerogatives.
   4. Continuously verify the compliance of corporate software with the new human oversight obligations under the AI Act to avoid joint liability between provider and user.

9. Update and rectification note (17 September 2026)

   A previous version of this article incorrectly referred to a non-existent appeal decision, Supreme Court jurisprudence on compensation for loss of chance to individual workers, and a reversal of the burden of proof deriving from the AI Act and GDPR. The text has been corrected to accurately reflect the only real act (Trib. Bologna ord. 31.12.2020), which ascertained algorithmic discrimination in the first instance, ordering the platform to compensate trade unions with 50,000 euros, applying the reversal of the burden of proof under Art. 28 of Legislative Decree 150/2011 and Art. 40 of Legislative Decree 198/2006.

References: Trib. Bologna ord. 31.12.2020D.Lgs. 198/2006D.Lgs. 150/2011Regolamento UE 2016/679 (GDPR)Regolamento UE 2024/1689 (AI Act)

Editorial oversight: Avv. Federico Papa·ICAMUpdated 17 September 2026Content drafted with AI support and subject to editorial source checks. Despite these controls, inaccuracies may remain: reports and rectification requests are welcome. Report a correction