Illustrazione di copertina: Trasparenza retributiva e valutazione del lavoro: i criteri del d.lgs. 96/2026 e gli strumenti automatizzati

Immagine creata con IAAI-generated image

Pay transparency and job evaluation: the criteria of Legislative Decree 96/2026 and automated tools

Legislative Decree No 96 of 7 May 2026, which implements Directive (EU) 2023/970 of the European Parliament and of the Council of 10 May 2023 on equal pay between men and women through pay transparency, has been in force since 7 June 2026 (Official Gazette (Gazzetta Ufficiale)). On 6 August 2026 the European Commission services published answers to frequently asked questions (FAQ) on the Directive, which devote considerable space to job evaluation criteria (European Commission). For the general framework we refer to what we set out on the occasion of International Equal Pay Day; here we examine the criteria laid down by the Italian rules for assessing work of equal value and the entrusting of that assessment to automated tools.

Article 4 of the Decree: the collective-agreement presumption and company evaluation systems

Article 4(1) provides that the application of a national collective agreement concluded by the comparatively most representative trade unions, “including the systems of job classification, grading and pay, constitutes a presumption of conformity with the principles of equal pay and transparency, without prejudice to proof of the existence of discriminatory individual pay treatment” (our translation) (Official Gazette (Gazzetta Ufficiale)).

Paragraph 4 identifies the classification systems laid down by collective agreements or by law as “the reference instrument for the purposes of comparison” and opens a space for company systems: “job classification and evaluation systems decided by the employer for the purposes of determining pay, supplementing what is provided for by the CCNL [national collective agreement], are also permitted, provided that they are characterised by objective and gender-neutral criteria”. Paragraph 6 allows the Minister of Labour and Social Policies to adopt guidance acts by 31 December 2026.

Article 6 requires employers to make accessible to workers “the criteria used to determine pay and pay levels, developed on the basis of Article 4” and regards the obligation as fulfilled, for those applying such a collective agreement, by reference to the criteria of the agreement and of any company-level agreements. Neither Article 4(1) nor Article 6(2) refers to the systems added by the employer: we therefore consider that the presumption and the simplified compliance cover what collective bargaining regulates, not supplementary company systems.

The evaluation criteria: comparing the Decree and the Directive

Under Article 4(4) of the Directive, the criteria for assessing whether workers are in a comparable situation “shall include skills, effort, responsibility and working conditions, and, if appropriate, any other factors which are relevant to the specific job or position”; they are also “agreed with workers’ representatives where such representatives exist”, and “relevant soft skills shall not be undervalued” (EUR-Lex).

Article 4(3) of the Decree provides instead that the assessment is to be carried out “on the basis of common, objective and gender-neutral criteria, which take into account skills, responsibilities and working conditions as well as any other factor relevant to the specific job or position”. Effort does not appear in the Decree’s list, and Article 4 reproduces neither the reference to criteria agreed with workers’ representatives nor the clause on soft skills.

The FAQ are explicit: the Directive “does not prescribe a specific tool”, but requires the use of “at least these four criteria”, described as a mandatory minimum requirement (European Commission). These are preliminary views of the services, not an official position, but they adhere to the text. In our view the employer should in any event include effort (impegno in the Italian version of the Directive) among the factors of its own system, relying on the clause concerning “any other relevant factor” and on an interpretation consistent with the Directive. A system based solely on the three criteria of the Decree risks failing to withstand comparison with the European benchmark.

The timetable: reporting from 7 June 2027 and implementing decrees still awaited

Article 9 concerns employers with at least one hundred employees: those employing at least one hundred and fifty collect the data by 7 June 2027, then every year from two hundred and fifty employees upwards and every three years below that threshold; those employing between one hundred and one hundred and forty-nine do so by 7 June 2031 and then every three years. A difference in average pay level “of at least 5 per cent in any category of workers”, not justified and not remedied within six months, triggers the joint pay assessment under Article 10, whose measures include “an analysis of the gender-neutral job evaluation and classification systems in place in the organisation or the establishment of such systems”.

Still awaited are the decrees of the Minister of Labour on the collection and processing of the data, to be adopted within ninety days after consulting the Italian Data Protection Authority (Garante per la protezione dei dati personali) (Article 9(4)), and those on the monitoring body, within one hundred and eighty days (Article 14(5)): the first deadline expired on 5 September 2026 and as at 10 October 2026 we had not found their publication in the Official Gazette.

Automated evaluation tools: the applicable framework beyond the Decree

Neither the Decree nor the Italian text of the Directive mentions algorithms, automated systems or artificial intelligence; the only reference is in the editorial notes to Article 6, which reproduce Article 1(1)(s) of Legislative Decree No 152 of 26 May 1997 on information concerning “automated decision-making or monitoring systems”.

Article 1-bis of that Decree concerns “fully automated” systems that provide indications relevant, among other things, to “the evaluation, performance and fulfilment of the contractual obligations of workers”; the information due includes the main parameters, “including performance evaluation mechanisms”, and the metrics, “as well as the potentially discriminatory impacts of those metrics”. Paragraph 8 excludes the obligations for “systems protected by industrial and trade secrets” (Normattiva).

In our view the link is necessary, even if it is not written down: job evaluation software that assigns scores to positions, or a tool that proposes pay bands and increases, falls among the evaluation systems decided by the employer permitted by Article 4(4) only if it is objective and neutral. Gender neutrality must be demonstrated on the model, not merely declared in the pay policy. Factors, weightings and training data are where the undervaluation of predominantly female work can be reproduced, as we have observed with regard to the new forms of discrimination in machine learning.

Article 22 of Regulation (EU) 2016/679 (GDPR) grants the right not to be subject to a decision “based solely on automated processing” producing legal effects or similarly significant effects (EUR-Lex), as we saw in the case of automated deactivations. Annex III, point 4(b), of Regulation (EU) 2024/1689, the AI Act, classifies as high-risk AI systems “intended to be used to make decisions affecting terms of work-related relationships” or to “monitor and evaluate the performance and behaviour of persons in such relationships” (EUR-Lex). Regulation (EU) 2026/1744, the Digital Omnibus on AI, has set 2 December 2027 as the date from which Chapter III, Sections 1, 2 and 3, applies to Annex III systems (EUR-Lex); from that date the employer acting as deployer must inform workers’ representatives and the affected workers before using the system (Article 26(7)). In our view an AI system that determines or proposes pay affects the “terms of work-related relationships”, whereas a spreadsheet with fixed scores remains outside the AI Act, but not outside Article 4 of Legislative Decree 96/2026.

On the national side, from 23 October 2026 Article 41 of Legislative Decree No 179 of 7 October 2026 requires employers using AI systems to ensure that final decisions on the establishment, modification or termination of the employment relationship, “including disciplinary measures and performance evaluations”, are not taken solely on the basis of automated processing, not even with the worker’s consent (paragraph 1), and declares null and void any dismissal and any decisions taken in breach (paragraph 4). Paragraph 3 contains two rules: before processing begins, the employer fulfils the information obligations under Article 1-bis of Legislative Decree 152/1997; the worker has the right to obtain, on request and through the intervention of a natural person, “an intelligible statement of the reasons for the decision concerning him or her”, including the possible influence of AI systems and “the main parameters considered” (Official Gazette (Gazzetta Ufficiale); cf. NicFab). For the overall framework we refer to our analysis of the Italian implementation of the AI Act. In our view a change in pay determined by the system modifies the relationship and falls within that limit, including the sanction of nullity.

Critical aspects: the scope of the collective-agreement presumption and the coordination of information notices

The presumption under Article 4(1) yields to proof of discriminatory individual treatment, but it shifts the centre of gravity of the evidence. The FAQ recall that, according to the Court of Justice (Cases C-400/93, Royal Copenhagen, and C-43/75, Defrenne), the principle also applies to pay set by collective bargaining, and employers covered by collective agreements are not excluded from the Directive. It follows, in our view, that the conformity of the collective agreement does not exempt the employer from verifying how it applies that agreement.

The second aspect concerns information notices. Industrial and trade secrets, which exclude the obligations under Article 1-bis, have no equivalent in Article 6 of Legislative Decree 96/2026: in our view secrecy about how the software works cannot be relied on against the obligation under Article 6, and those using a third-party tool must obtain from the supplier enough to describe its criteria. Article 41(3) of Legislative Decree 179/2026 requires the obligations under Article 1-bis to be fulfilled before processing begins, but does not say whether the reference extends the information notice to AI systems that are not fully automated or merely recalls it for the cases already covered; we consider it prudent to provide it for every system used in the decisions referred to in Article 41. The real risk is a generic description of the criteria, while the actual functioning of the tool remains opaque.

Conclusions

Legislative Decree 96/2026 entrusts the neutrality of evaluation to collective agreements and to objective and neutral company systems. Three checks remain with the employer: including effort among the evaluation factors, as the Directive requires; documenting automated tools so as to be able to set out their criteria under Article 6; coordinating with pay transparency the information notices required by Legislative Decree 152/1997, the GDPR, Legislative Decree 179/2026 and, from 2 December 2027, the AI Act. Pending the implementing decrees and any guidance acts, for employers with at least one hundred and fifty employees the first data collection is set for 7 June 2027: the review of evaluation systems should precede that date.

Author: Valentina Grazia Sapuppo


AI AnthropoCosmic In evidenzaAI AnthropoCosmicUn progetto internazionale per un’IA a servizio dell’Uomo, dell’Ambiente e del Cosmo, che mette al centro la dignità della persona nella progettazione dei sistemi. Leggi l’articoloAI Open Mind AI AnthropoCosmic FeaturedAI AnthropoCosmicAn international project for an AI at the service of humanity, the environment and the cosmos, placing human dignity at the centre of system design. Read the articleAI Open Mind Agentic AI In evidenzaAgentic AILimiti prima dell’azione, evidenze durante, responsabilità dopo. Il volume di Nicola Fabiano sulla governance dei sistemi agentici, con la prefazione di Antonino Caffo.Capitolo 16 a cura dell’Avv. Valentina Grazia SapuppoLeggi l’articolo Agentic AI FeaturedAgentic AILimits before the action, evidence during, responsibility afterwards. Nicola Fabiano’s book on the governance of agentic systems, with a preface by Antonino Caffo.Chapter 16 by Valentina Grazia SapuppoRead the article
Intervista Radio Radio IntervistaLegge e colossi del digitaleIl patteggiamento di Meta sui minori non è una condanna. Stati Uniti ed Europa seguono strade opposte, e sugli agenti di IA resta aperta la domanda su chi risponde.Un Giorno Speciale su Radio Radio, con Alessio De Paolis · audio dal minuto 2:26:00Ascolta l’intervistaGuarda il videoLeggi l’articolo
Digital Omnibus ContributoIl Digital Omnibus cambia l’AI ActNuove scadenze per i sistemi ad alto rischio e un chiarimento sull’obbligo di AI literacy: più tempo per adeguarsi, nessuno sconto sulla preparazione di persone e processi.Articolo scritto per il blog di SkillaLeggi su Skilla
Digeat Festival 2026 SpeakerDigeat Festival 2026Valentina Grazia Sapuppo tra i relatori del festival dedicato a protezione dei dati, archivi digitali e regole del futuro. Interviene sul tema «Le regole dell’IA: nuove leggi o principi del diritto?».Venerdì 6 novembre 2026, ore 16:30, Ex Convitto Palmieri, LecceL’interventoLa scheda relatriceIl festival
AI AnthropoCosmic 2026AI AnthropoCosmic 2026Valentina Grazia Sapuppo nel progetto dell’Università Pontificia Salesiana su Persona, Ambiente e Cosmo: moderazione della sessione mattutina del Convegno finale e intervento negli AI Laboratori del Domani su commercio elettronico e IA.14 novembre 2026, online · 28 novembre 2026, Università Pontificia Salesiana, RomaIl convegnoL’incontroIl contributoIl comitato scientificoLa relatriceIl progetto
Interview Radio Radio InterviewLaw and the digital giantsThe Meta settlement on minors is not a conviction. The United States and Europe take opposite paths, and on AI agents the question of who answers remains open.Un Giorno Speciale on Radio Radio, with Alessio De Paolis · audio from 2:26:00 · in ItalianListen to the interviewWatch the videoRead the article
Digital Omnibus ContributionThe Digital Omnibus reshapes the AI ActNew deadlines for high-risk systems and a clarification on the AI literacy duty: more time to comply, no discount on preparing people and processes.Article written for the Skilla blog, in ItalianRead on Skilla
Digeat Festival 2026 SpeakerDigeat Festival 2026Valentina Grazia Sapuppo among the speakers of the festival on data protection, digital archives and the rules of the future. She takes part in the panel «The rules of AI: new laws or principles of law?».Friday 6 November 2026, 16:30, Ex Convitto Palmieri, LecceThe panelSpeaker profileThe festival
AI AnthropoCosmic 2026AI AnthropoCosmic 2026Valentina Grazia Sapuppo in the project of the Università Pontificia Salesiana on Person, Environment and Cosmos: moderator of the morning session of the closing conference and speaker at the AI Laboratori del Domani on e-commerce and AI.14 November 2026, online · 28 November 2026, Università Pontificia Salesiana, Rome · sessions held in ItalianThe conferenceThe sessionThe contributionThe scientific committeeSpeakerThe project