On 8 October 2026, Legislative Decree No. 179 of 7 October 2026 was published in the Official Gazette, aligning Italian legislation with Regulation (EU) 2024/1689 on artificial intelligence (AI Act).
The Decree will enter into force on 23 October 2026 and introduces, among other provisions, specific rules governing the use of artificial intelligence systems in employment relationships, as well as in the field of occupational health and safety.
For employers, the most relevant developments concern mandatory human oversight of employment-related decisions, information obligations towards employees, and the requirement to update workplace risk assessments where AI affects work organisation or occupational health and safety.
Employment-Related Decisions: Mandatory Human Oversight
The employer using AI systems in decision-making processes relating to the employment relationship must ensure that final decisions are not made solely on the basis of automated processing. In particular, the final decision must be reserved for a natural person with effective and independent decision-making authority.
This protection applies to decisions concerning:
- the establishment of the employment relationship;
- changes to the employment relationship;
- termination of the employment relationship;
- disciplinary measures;
- performance assessments.
The Decree clarifies that candidate screening and selection activities, including where they result in a candidate not being admitted to the subsequent stages of the recruitment process, do not qualify as final decisions concerning the establishment of the employment relationship. In any event, the employee’s consent does not allow any derogation from the prohibition on final decisions based exclusively on automated processing. It should be noted that this exclusion relates specifically to the classification of the selection process as a final decision, without prejudice to the other obligations laid down by the AI Act, including compliance with the requirements concerning the transparency of decision-making systems, anti-discrimination legislation and data protection laws.
Before the system is used in the relevant decision-making or monitoring process, the employer must in any event comply with the information obligations concerning the use of fully automated decision-making or monitoring systems already provided for under Article 1-bis of Legislative Decree No. 152/1997. Such information must specify, among other things, the aspects of the employment relationship affected, the purposes and objectives of the system, its operating logic, the categories of data and main parameters used, as well as the monitoring measures adopted. Employees may also request an intelligible explanation of any decision concerning them, provided through the intervention of a natural person and including information on the potential impact of the AI system on the decision-making process and the main parameters taken into account.
The Decree expressly provides that any dismissal carried out in breach of the human oversight requirement is null and void, as are any other decisions taken by the employer in breach of the same human oversight requirement.
AI and occupational health and safety: updating risk assessment and training
When AI systems affect work organisation, production rates, the way work is performed or decision-making processes relevant to occupational health and safety, their use must be taken into account as part of the risk assessment required under Article 28 of Legislative Decree No. 81/2008.
Employers must also ensure that employees receive adequate information and training on the specific risks associated with the use of AI systems and on the prevention and protection measures adopted.
Il Decreto individua un sistema di vigilanza che coinvolge l’ACN, l’AgID e, nei rispettivi settori, le altre autorità competenti, ferme restando le attribuzioni del Garante per la protezione dei dati personali. Il controllo può essere avviato a seguito della segnalazione di incidenti gravi, di reclami non manifestamente infondati, su segnalazione di altre autorità o d’ufficio dall’ACN. Nell’ambito delle prove in condizioni reali, sono inoltre previsti controlli a distanza o in loco, anche senza preavviso. In tale contesto, per le imprese assume particolare importanza predisporre adeguati presidi interni che consentano di documentare l’utilizzo dei sistemi di AI, gestire tempestivamente eventuali contestazioni e assicurare una supervisione umana effettiva dei processi decisionali.
Supervision, inspections and sanctions
The Decree establishes a supervisory framework involving ACN, AgID and, within their respective areas of competence, the other competent authorities, without prejudice to the powers of the Italian Data Protection Authority. Supervisory action may be initiated following reports of serious incidents, complaints that are not manifestly unfounded, reports from other authorities, or on ACN’s own initiative. In the context of testing under real-world conditions, remote or on-site inspections, including unannounced inspections, are also provided for. Against this background, it is particularly important for companies to implement appropriate internal safeguards to document the use of AI systems, promptly address any challenges or complaints, and ensure effective human oversight of decision-making processes.
The Decree coordinates the national sanctions framework with that provided for under the AI Act. For the most serious infringements covered by the AI Act, administrative fines may amount to up to €35 million or, if higher, 7% of the worldwide annual turnover for the preceding financial year. For less serious or less harmful infringements, the competent authorities may instead impose non-pecuniary measures, including an order to bring the infringement to an end and refrain from repeating it.
Practical Guidance for Employers
In view of the entry into force of the Decree, companies that already use or intend to introduce AI tools in HR processes should therefore, as part of their AI governance obligations, also carry out a specific internal review aimed at:
- mapping the AI systems used in recruitment, personnel management, performance assessment, disciplinary and termination processes, including systems provided by external service providers;
- identifying the decisions for which effective and independent human intervention must be ensured;
- reviewing and, where necessary, updating the information provided to employees and agreements with recruitment service providers;
- establishing procedures to ensure human oversight, identify the persons responsible and provide intelligible explanations of AI-assisted decisions;
- assessing the impact of AI systems on the workplace risk assessment document and on employee information and training programmes.