Navigating uncharted waters : select practical considerations in radiology AI compliance with the EU AI Act

Potočnik, Jaka (author), Fujs, Damjan (author)

Artificial intelligence (AI) is transforming traditional medicine, particularly in radiology. Its integration across patient care stages has made it increasingly ubiquitous. The European Union’s (EU) AI Act will additionally regulate AI-enabled solutions within the EU. However, without standardized guidelines, the Act’s flexibility poses practical challenges for providers and deployers, leading to inconsistencies in meeting requirements for high-risk systems like radiology AI, potentially impacting patients’ fundamental rights and safety. Many healthcare providers in the (EU) have already integrated AI-enabled medical devices or systems into their radiology workflows. While their conformity with the Medical Device Regulation (MDR)$^1$ was primarily assessed by a notified body, all of these will fall under the scope of the AI Act$^2$.This new EU regulation complements the existing MDR and has been fully in force since August 2025 for general-purpose AI and foundation models, with a transition period of up to two years for high-risk AI systems, such as medical devices. Most applications of AI in radiology are likely to be classified as high-risk due to processing personal health data, influencing human experts’ decision-making, impacting patient outcomes and quality of life. As such, these systems must uphold the rights specified in the Charter$^3$, including the right to protection of personal data (Article 8), nondiscrimination (Article 21), equality between men and women (Article 23), and access to health care (Article 35). Consequently, more stringent requirements apply to high-risk AI systems$^4$.

• Language: English
• Keywords: computer science, radiology, medicine, artificial intelligence, risk management, medical devices, science
• Work type: Article
• Typology: 1.03 - Other scientific articles
• Organization: FRI - Faculty of Computer and Information Science
• Publication status: Published
• Publication version: Version of Record
• Year: 2025
• Number of pages: 3 str.
• Numbering: Vol. 8, art. 630
• PID: 20.500.12556/RUL-175599
• UDC: 004.8:615.849
• ISSN on article: 2398-6352
• DOI: 10.1038/s41746-025-02094-z
• COBISS.SI-ID: 254951427
• Publication date in RUL: 05.11.2025

Record is a part of a journal

• Title: NPJ digital medicine
• Shortened title: npj digit. med.
• Publisher: Springer Nature in partnership with the Scripps Research Translational Institute
• ISSN: 2398-6352
• COBISS.SI-ID: 529714201

Licences

• License: CC BY 4.0, Creative Commons Attribution 4.0 International
• Link: http://creativecommons.org/licenses/by/4.0/
• Description: This is the standard Creative Commons license that gives others maximum freedom to do what they want with the work as long as they credit the author.

Secondary language

• Language: Slovenian
• Keywords: računalniška znanost, radiologija, medicina, umetna inteligenca, obvladovanje tveganj, medicinski pripomočki, znanost