EU AI Act use-case guide · Last verified 2026-01-15High risk

EU AI Act for AI clinical decision support in Healthcare & Medical Technology

CDSS that materially shapes clinician decisions is Annex III §1 high-risk and differs from acuity-scoring triage by recommending an action.

Preliminary risk score 89/100Annex III, §1Preliminary summary · Not legal advice
AI clinical decision supportCDSS EU AI Actdiagnostic recommendation AIAnnex III §1 CDSStreatment path AI

Risk level

AI clinical decision support maps to a high-risk Annex III category, so the obligations below apply in full.

Annex III anchor

Annex III, §1

Score basis

A preliminary 89/100 based on the type of decision the system influences and how it is deployed in Healthcare & Medical Technology.

Provider obligations

What the provider (developer) must do

Art. 10

Data governance across patient cohorts and guidelines

EUR-Lex
Art. 15

Accuracy, robustness, cybersecurity

EUR-Lex
Art. 11

Technical documentation of clinical evidence

EUR-Lex

Deployer obligations

What you must do as the deployer

Art. 14

Treating clinician oversight on each recommendation

EUR-Lex
Art. 26

Audit trail of recommendations vs final clinical decision

EUR-Lex

Deployment

How AI clinical decision support shows up in Healthcare & Medical Technology

Typical contexts

Hospital EHR-integrated CDSS panelSpecialty clinic second-opinion tooling

Signals it's in play

  • CDSS recommendation
  • Treatment-path suggestion
  • Diagnostic differential

Recommendations

  • Treating clinician decides
  • Track override rate vs recommendation
  • Periodic clinical outcome audit

Watch-outs

  • Anchoring on AI suggestion
  • Outdated guideline knowledge
  • Uncalibrated alert thresholds

FAQ

EU AI Act questions about AI clinical decision support

Is AI clinical decision support high-risk under the EU AI Act?

AI clinical decision support maps to Annex III, §1, which the EU AI Act treats as high-risk. In practice it is assessed as High risk, and the obligations below apply to providers and deployers.

Which EU AI Act articles apply to AI clinical decision support?

The obligations that typically apply are Art. 10 — data governance across patient cohorts and guidelines; Art. 15 — accuracy, robustness, cybersecurity; Art. 11 — technical documentation of clinical evidence; Art. 14 — treating clinician oversight on each recommendation; Art. 26 — audit trail of recommendations vs final clinical decision. Providers (developers) carry the technical duties; deployers (operators) carry the use, oversight, and transparency duties.

Who is responsible — the provider or the deployer of AI clinical decision support?

Both. Providers owe the technical obligations such as Art. 10, Art. 15, Art. 11. Deployers owe Art. 14, Art. 26. The split matters for procurement and vendor agreements in Healthcare & Medical Technology.

What should you watch out for with AI clinical decision support?

Common failure modes include: Anchoring on AI suggestion; Outdated guideline knowledge; Uncalibrated alert thresholds. Mitigations typically start with Treating clinician decides and Track override rate vs recommendation.

Where does AI clinical decision support typically appear in Healthcare & Medical Technology?

Typical deployment contexts include Hospital EHR-integrated CDSS panel and Specialty clinic second-opinion tooling. Before deploying, confirm whether the specific use triggers the high-risk obligations listed above.

Sources

Citations & further reading

Related

More AI use cases in Healthcare & Medical Technology

Limited risk38/100

Customer support chatbot

Automates customer conversations and support triage.

Read the guide
High risk91/100

Medical triage AI

Supports triage, diagnosis, or prioritization in healthcare settings.

Read the guide
Prohibited risk96/100

Biometric identification

Identifies or verifies people using biometric characteristics.

Read the guide
High risk60/100

Biometric access control (workforce)

Employee or patient biometric authentication for premises or systems.

Read the guide
Limited risk47/100

AI on-call staff rostering

Allocates on-call shifts to clinical or operational staff based on demand/availability.

Read the guide
High risk93/100

AI medical image analysis

Analyses radiology/pathology imaging for diagnostic decision support.

Read the guide
High risk95/100

AI surgical robot assistant

Provides real-time guidance or autonomous sub-steps during surgery.

Read the guide
High risk83/100

Continuous patient monitoring AI

Continuously monitors inpatient vitals and raises early-warning scores.

Read the guide
High risk67/100

AI clinical-trial matching

Suggests clinical-trial enrolment based on patient profile and trial criteria.

Read the guide
High risk72/100

AI insurance claim triage

Routes or prioritises claims for fast-track, manual review, or SIU escalation.

Read the guide
High risk78/100

AI content moderation

AI that flags, removes, or ranks user-generated content.

Read the guide
Minimal risk18/100

AI retail demand forecasting

Forecasts demand to drive inventory and procurement decisions.

Read the guide
High risk94/100

AI emergency call prioritisation

Scores and prioritises inbound emergency calls for first-responder dispatch.

Read the guide
Limited risk46/100

AI vendor credentialing

Onboarding AI that scores vendor documents, KYB data, and risk signals.

Read the guide
Limited risk35/100

AI document summarisation

Generates concise summaries of long regulatory or contractual documents.

Read the guide

Explore

More industry guides

Describe your exact system, get a personalised read

The guide above is a general baseline for AI clinical decision support. The free Risk Scanner maps your specific implementation and surfaces hidden compliance blind spots.

Open the Risk Scanner

Preliminary EU AI Act clarity summary. Not legal advice.