Algorithmic dismissal and the right to explanation: nullity for software opacity

1. The fact

   The case arises from the widespread use of automated management systems in the delivery and logistics sectors, as reported by La Stampa and Wired Italia. In the Foodinho case in Palermo (Trib. Palermo ord. 20.6.2023), the court addressed the operation of the Jarvis algorithm, which managed rider shifts. In a separate proceeding pursuant to Article 28 of the Workers' Statute against Uber Eats, also in Palermo, the judge rejected the company's defense based on trade secrets and ordered the disclosure of the ranking criteria. Similarly, the Court of Bologna (order 31.12.2020) had previously censured Deliveroo's Frank algorithm due to its discriminatory impact on workers exercising the right to strike.

2. The rules at play

   1. EU Regulation 2016/679 (GDPR): Article 22 guarantees the right not to be subject to decisions based solely on automated processing and requires the controller to provide meaningful information about the logic involved.
   2. AI Act (Regulation EU 2024/1689): Article 86 introduces the right to an explanation for decisions taken by high-risk systems, including those used for workforce management, requiring clarity on decision parameters.
   3. Transparency Decree (Legislative Decree 104/2022): Article 1-bis mandates analytical disclosure regarding automated monitoring and decision-making systems, the violation of which compromises the transparency of the relationship.
   4. Law 604/1966: provides that a dismissal lacking motivation or supported only by generic grounds is punishable by nullity or ineffectiveness, as it prevents the worker from exercising the right of defense.

3. What case law says

   Trial courts, with an emerging orientation, have clarified that trade secrets cannot prevail over the worker's fundamental right to understand the criteria governing their employment and career. Judges have established that the employer bears the burden of translating algorithmic language into clear legal and factual terms, under penalty of the termination's illegality. Furthermore, case law has consolidated the principle that an algorithm failing to distinguish reasons for worker unavailability, such as illness or strike participation, constitutes indirect discrimination. European courts have recently reiterated that algorithmic profiling producing significant legal effects requires a degree of explainability that enables the data subject to challenge the substance of the decision, rather than merely acknowledging its formal existence.

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. Preventive Algorithmic Audit: corporate counsel must advise reviewing software procurement contracts to include indemnity clauses and technical disclosure requirements for suppliers.
   2. Documenting Human Intervention: it is essential to establish documentary evidence demonstrating genuine human review of AI outputs, countering allegations of purely automated decision-making.
   3. Analytical Disclosure: a generic contractual clause is insufficient; employers must provide a dynamic disclosure specifying input data and its weight in performance evaluations.
   4. Litigation Management: in court, defense strategy must focus on expert evidence and technical reports translating system logs into factual motivations compliant with Law 604/1966.

9. Update and rectification note (17 September 2026)

   The previous version of this article incorrectly attributed the Jarvis algorithm to Uber Eats and cited non-existent rulings on the nullity of collective dismissals due to AI opacity, as well as an untraceable case of dismissal for low performance. The text has been corrected to separate the Foodinho case (Trib. Palermo ord. 20.6.2023, Jarvis algorithm) from the Art. 28 proceeding against Uber Eats (Trib. Palermo), and to clarify that the nullity of algorithmic dismissal is currently a doctrinal theory. The overcoming of trade secrets has also been reclassified as an emerging orientation, integrating the correct references, including the Trib. Bologna order 31.12.2020 on Deliveroo.

References: Regolamento UE 2016/679 (GDPR)Regolamento UE 2024/1689 (AI Act)D.Lgs. 104/2022Legge 604/1966Articolo 28 Legge 300/1970Trib. Palermo ord. 20.6.2023Trib. Bologna 31.12.2020

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