Dal non discriminare al capacitare.Gestione algoritmica, disabilità e accomodamento ragionevole nell’era dell’AI Act

This article examines the relationship between algorithmic work management, disability and the employer’s duty of reasonable accommodation in the era of Regulation (EU) 2024/1689 (the AI Act). Proceeding from the finding that the Regulation is structurallyinsufficient, in itself, to establish a positive obligation on the part of the employer-deployer to assess algorithmic impact, the author reconstructs a composite duty of algorithmic reasonable accommodation, grounded in Article 5-bis of Law No 104/1992 as amended by Legislative Decree No 62/2024, and in Articles 28–29 of Legislative Decree No 81/2008. The obligations arising under the AI Act –monitoring and suspension duties under Article 26(5), and the Fundamental Rights Impact Assessment under Article 27 –interact with this normative framework in a differentiated manner depending on the deployer’s position. The obligation has a negative dimension, prohibiting the use of systems whose differential effects on workers with disabilities lack objective justification; and an active dimension, requiring a proactive assessment of whether available AI technologies can enhance the worker’s functional capabilities. The applicable proportionality test is interpreted as a dynamic criterion, calibrated on the principle of adaptation to technical progress under Article 6(2)(e) of Directive 89/391/EEC. Within the identified structural constraints –the procedural gap in the composition of the multidisciplinary assessment team, the inadequacy of anti-discrimination instruments in addressing intersectional algorithmic discrimination, and the asymmetry of liability between provider and deployer –the shift from non-discrimination to enablement can already be derived, through interpretation, from the existing body of law.