Compliance
The EU AI Act for healthcare: high-risk AI explained
Short answer
The EU AI Act treats most AI in medical devices, and some health-related uses such as emergency triage and health insurance pricing, as high-risk. Providers must run risk management, data governance, documentation and human oversight. After the Digital Omnibus, these duties apply from 2 December 2027 for stand-alone systems and 2 August 2028 for AI in regulated products.
- Author
- Kumbirai Muhwati
- Published
- Last reviewed
- Reading time
- 4 min
Does it apply to a US company?
It can. Under Article 2, the Act applies to providers that place AI systems on the EU market or put them into service in the EU, wherever the provider is based. It also applies to providers and deployers outside the EU when the output of their AI system is used in the EU. A US vendor selling to European hospitals, or a US research program whose AI outputs are used for participants in the EU, should assess its exposure.
The risk tiers
- Prohibited practices (Article 5), such as manipulative techniques that cause significant harm. These bans have applied since 2 February 2025.
- High-risk systems (Article 6, Annexes I and III), which carry the main obligations described below.
- Transparency obligations (Article 50), for example telling people when they are interacting with an AI system.
- Minimal-risk systems, which carry no specific obligations.
Separate rules apply to general-purpose AI models, such as large language models, and have applied since 2 August 2025.
Why most medical-device AI is high-risk
Under Article 6(1), an AI system is high-risk if it is a product, or a safety component of a product, covered by the EU legislation listed in Annex I, and that product must undergo a third-party conformity assessment. The Medical Devices Regulation (MDR) and the In Vitro Diagnostic Medical Devices Regulation (IVDR) are both in Annex I. Medical device software in MDR class IIa or higher, and IVD software in class B or higher, generally needs a notified body, so AI in those devices is usually high-risk.
Annex III uses that touch health
- Evaluating whether people are eligible for essential public assistance benefits and services, including healthcare services.
- Risk assessment and pricing for life and health insurance.
- Evaluating and classifying emergency calls, dispatching emergency services including medical aid, and emergency healthcare patient triage.
Annex III also covers areas such as employment and biometrics, which can apply to healthcare organizations in their role as employers.
What providers and deployers must do
Providers, the organizations that develop a high-risk system and place it on the market, must put in place:
- A risk management system that runs for the whole lifecycle (Article 9).
- Data governance for training, validation and test data (Article 10).
- Technical documentation (Article 11) and automatic logging of events (Article 12).
- Clear instructions and information for deployers (Article 13).
- Design for effective human oversight (Article 14).
- Appropriate accuracy, robustness and cybersecurity (Article 15).
- A quality management system (Article 17) and a conformity assessment before the system is placed on the market.
- Post-market monitoring (Article 72) and reporting of serious incidents (Article 73).
Deployers, such as a hospital using a high-risk system, must use it according to its instructions, assign competent people to oversee it, monitor it and keep its logs (Article 26). Public bodies and some private deployers, including those using AI for life and health insurance pricing, must also carry out a fundamental rights impact assessment (Article 27).
The timeline after the Digital Omnibus
- 1 August 2024: the AI Act entered into force.
- 2 February 2025: prohibitions and AI literacy obligations began to apply.
- 2 August 2025: obligations for general-purpose AI models began to apply.
- 2 December 2027: high-risk obligations apply to stand-alone systems listed in Annex III (moved from 2 August 2026 by the Digital Omnibus on AI, Regulation (EU) 2026/1744).
- 2 August 2028: high-risk obligations apply to AI in products covered by Annex I, including medical devices under the MDR and IVDR (moved from 2 August 2027).
How alignment differs from conformity assessment
Conformity assessment is the formal procedure a provider completes before placing a high-risk system on the market. Depending on the system, it is done internally or by a notified body, and it leads to an EU declaration of conformity and CE marking. Alignment means applying the Act’s practices, such as risk management, documentation and human oversight, before those obligations formally apply. It is not a claim of conformity or certification.
Frequently asked questions
- Is all healthcare AI high-risk under the EU AI Act?
- No. AI that is, or is part of, a medical device needing a notified body is usually high-risk, as are certain Annex III uses such as emergency triage and health insurance pricing. Many administrative tools are not high-risk, although transparency rules may still apply.
- Does the EU AI Act replace the MDR?
- No. Both apply to AI in medical devices. The AI Act lets providers include its required information in the technical documentation they already prepare under the MDR or IVDR.
- When do the high-risk obligations apply?
- From 2 December 2027 for stand-alone systems listed in Annex III, and from 2 August 2028 for AI in products covered by Annex I, such as medical devices.
- What is the Digital Omnibus on AI?
- Regulation (EU) 2026/1744, which amended the AI Act in 2026. Among other changes, it postponed the dates on which the high-risk obligations apply.
Sources
- Regulation (EU) 2024/1689, the AI Act (EUR-Lex)
- Regulation (EU) 2026/1744, the Digital Omnibus on AI (EUR-Lex)
- Regulation (EU) 2017/745, the Medical Devices Regulation (EUR-Lex)
General information, not legal or medical advice.