ITALY Trends and Developments Contributed by: Michela Bani, Alessandro Paone and Giacomo Bertelli, NIUS Legal and HR Solutions
AI tends to concentrate and overlap, within a single technological infrastructure, the managerial power, the organisational power and the power of control of the employer. Internal investigations, AI and legal risk qualification Under the European and national approach, the use of AI systems in internal investigations falls into the cat - egory of “high-risk” jobs, as it is likely to directly affect the fundamental rights and freedoms of workers. This category includes systems intended for: • the automated analysis of electronic communica - tions (email, chat, messaging systems); • the behavioural profiling of employees; • the identification of anomalies, suspicions of fraud or disciplinary violations; and • support decisions in the evaluation of individual conduct. Regulation (EU) 2024/1689 (AI Act) and Italian Law No 132 of 23 September 2025 qualify these uses as being deserving of enhanced safeguards, based on obligations of transparency, prior risk assessment, traceability and human control. It follows that AI cannot be considered a mere neu - tral technical tool, but a structural component of the organisation of work, with significant legal and trade union implications. Human centrality and prohibition of automated decision-making One of the cornerstones of the discipline is the prin - ciple of the centrality of human intervention. In line with the European anthropocentric model, Law No 132/2025 states that AI systems must be developed and used in compliance with human decision-mak - ing autonomy, ensuring the possibility of supervision, intervention and correction. Transposed into the context of internal investigations, this principle means that: • AI can support data analysis and the identification of relevant patterns; • AI cannot, however, automatically determine the disciplinary liability of a worker; and
• the final assessment of the facts, the legal clas - sification of the conduct and the adoption of the consequent measures must remain the exclusive prerogative of a natural person. Any form of automation of the disciplinary decision would expose the company to serious accusations of illegitimacy, from the point of view of both employ - ment (violation of Article 7 of the Labour Statute) and the protection of personal data. Transparency, disclosure obligations and the right to explanation The use of AI in internal investigations has a direct impact on the employer’s disclosure obligations. The regulatory framework consists of Legislative Decree 152/1997, as amended by Legislative Decree 104/2022, the GDPR and Law No 132/2025, and requires enhanced transparency. The employer is required to inform workers, in a clear, accessible and documented manner, about: • the existence of AI systems used in the organisa - tion and management of the employment relation - ship; • the purposes pursued (including the prevention and detection of offences); • the operating logic of the systems; • the categories of data processed and the main parameters used; • human control measures and correction processes; and • the possible effects of algorithmic analyses on the worker’s position. This is accompanied by the worker’s right to obtain a comprehensible explanation of the decisions or assessments relevant to said worker – a principle now recognised both by European legislation and by the jurisprudence of the Court of Justice. In the context of internal investigations, this right takes on a particularly delicate importance: the use of opaque or inexplicable algorithmic outputs risks compromising the right of defence and determining the illegality of the entire procedure.
256 CHAMBERS.COM
Powered by FlippingBook