The findings of an audit by the Inspectorate-General for Health Activities (IGAS) indicate that the implementation of artificial intelligence (AI) within the National Health Service (SNS) is at an early and inconsistent stage.
The audit aimed to assess whether the SNS hospital entities—comprising 39 local health units and three oncology institutes—ensure that this technology enhances healthcare quality, protects user rights, and promotes an efficient, ethical, and secure management of public resources.
IGAS concluded that only 12 of the audited entities have a defined strategy for AI, with 10 in the process of development, while 14 have no strategy at all.
Regarding expertise and specialized knowledge, the inspection found that 14 entities have AI training plans, but 60% do not promote digital literacy in this area for health professionals or users.
“There is a noted need for greater capacity building for the safe and ethical use of artificial intelligence,” the report states.
Another finding shows that 21 entities have adopted measures to ensure rights such as privacy and non-discrimination, but only three inform users when interacting with AI, for example, through virtual assistants.
“Ethically, while most entities take measures to ensure rights like privacy, there is little transparency: only three inform users that they are interacting with an AI system,” IGAS warns, noting a lack of systematic evaluation of AI’s benefits and risks.
In terms of security, the audit found that 16 SNS entities have specific cybersecurity measures for AI, 15 have risk management plans, but many are still under development, with significant concerns about cyberattack risks and unauthorized access.
Concerning financing, there is an external dependency with limited self-financing capacity, as various local health units resort to sources like the Recovery and Resilience Plan (RRP), the Innovation and Digital Transition Program (Compete 2020), and the Horizon Europe program, IGAS reports.
“The audit showed that AI in the SNS is in its initial phase, with some progress in certain entities, but there is a lack of coordination and investment in critical areas,” the document highlights, suggesting that the findings may serve “as an additional tool to guide the ethical, safe, and efficient implementation of artificial intelligence in healthcare.”
Given it is an exploratory audit focused on the period of 2024 and the first quarter of 2025, IGAS did not issue formal recommendations to the local health units and oncology institutes.
