Context: Potential advantages and risks of AI and the Italian regulatory framework
Artificial Intelligence (AI) is reshaping healthcare by addressing systemic challenges and enhancing care delivery. Key potential benefits include improved diagnostic accuracy, reduced medical errors, increased efficiency through task automation and expanded access via telemedicine and decision-support tools. AI can also strengthen public health by enabling data-driven epidemic forecasting and policy evaluation.
However, economist Daron Acemoglu and colleagues caution that AI’s impact depends on its trajectory. If used primarily for automation (“so-so automation”), it may replace human labour without boosting productivity, leading to stagnating job growth and rising inequality. In contrast, “reinstating AI” – which complements human skills – can foster inclusive growth and better service quality.
Four major risks emerge:
- Labour substitution: AI may reduce demand for healthcare workers, especially if focused on replacing rather than augmenting tasks.
- Excessive automation: Economic incentives might drive inefficient adoption of AI, prioritizing cost-cutting over clinical improvement.
- Inequality: Automation could widen gaps between high- and low-skilled workers and between urban and rural patients.
- Skills mismatch: Rapid technological shifts may outpace training systems, leading to underuse or resistance among healthcare professionals.
To maximize AI’s potential, healthcare systems should invest in complementary innovations, equitable access and workforce development, ensuring technology enhances rather than undermines care.
In Italy, the approval of Law No. 132 of 23 September 2025 (“Provisions and Delegations to the Government on Artificial Intelligence”) represents the country’s commitment to integrating AI responsibility into the healthcare system and aligning with EU Regulation 2024/1689 (the AI Act). The law introduces measures expected to promote ethical and responsible use of AI in healthcare: a ban on discriminatory applications, protection of privacy and the right to information, and a requirement that AI support rather than replace human clinical judgment (for example, in radiology). It also establishes a national AI platform, managed by Agenas, integrated with the Electronic Health Record (Gazzetta Ufficiale, 2025).
Main measures
We report here in more detail some articles of the law:
- Article 7 establishes that AI technologies must support medical professionals in prevention, diagnosis and treatment, without replacing human clinical judgment. An example is in radiology: AI applications cannot completely replace human reading. The law also places special emphasis on people with disabilities, promoting AI systems that enhance autonomy, accessibility and social inclusion.
- Articles 8 and 9 recognize the public interest in the use of health data for research, allowing the processing of sensitive and pseudonymized data under strict safeguards, and introduce provisions for experimental AI projects under Ministry of Health oversight.
- Article 10 mandates the creation of a national AI platform, managed by Agenas, designed to support healthcare professionals with clinical guidance and provide citizens with interactive services. This platform will be integrated with the Electronic Health Record, leveraging intelligent data analysis while fully respecting GDPR and supervised by relevant authorities. However, it is likely that the implementations of these applications will take time and may be heterogeneous across territories.