The Complete Guide to Using AI in the Healthcare Industry in Belgium in 2025

Belgium in 2025 sits at a practical inflection point: a federal coalition and an emerging national AI and data strategy aim to harness AI for better diagnostics, telemonitoring and more joined-up care while keeping cybersecurity front and centre, and the Belgian Integrated Health Record (BIHR) is being rolled into the European Health Data Space to reach full interoperability by 2026–2027; read more on Belgium's roadmap in the Chambers Digital Healthcare 2025 review and the European Commission's overview of AI in healthcare.

With strong regulatory guardrails (AI Act, EHDS, NIS2) and lingering reimbursement bottlenecks, clinicians and teams need accessible, job-ready training - practical courses like the AI Essentials for Work bootcamp help bridge the “trust vs.

understanding” gap so AI tools actually improve care delivery.

Bootcamp	Length	Early bird cost	Registration
AI Essentials for Work bootcamp - course syllabus	15 Weeks	$3,582	Register for the AI Essentials for Work bootcamp

"AI is no longer a future promise, it's a present force." - HLTH Europe 2025

The Belgian AI strategy is deliberately practical and multi‑layered: at federal level it frames three strategic pillars - boosting technological impact and data governance, growing human capital and economic uptake, and securing ethical, resilient AI - while recognising that Belgium's federated politics means real delivery lives with regions and communities; see the European Commission Belgium AI Strategy Report and regional breakdown.

The new 2025–2029 Coalition Agreement commits to a national data and AI strategy developed with the Data Protection Authority, nominates BIPT as the national supervisor under the EU AI Act, and signals living labs and sandboxes for controlled experimentation (including law‑enforcement pilots) so innovation can be tested without risking patient or citizen trust - see the Belgium 2025–2029 Coalition Agreement and national data and AI strategy.

Practically, that translates into sizeable, targeted regional programmes (Flanders' AI action plan and research programme, Wallonia's DigitalWallonia4.ai and ARIAC/TRAIL projects, and Innoviris funding in Brussels), coordinated open‑data and cloud‑first ambitions, and a clear health focus: interoperability, BIHR/EHDS alignment and safe secondary use of data.

The

“so what?”

is concrete: rather than top‑down proclamations, Belgium pairs ethical guardrails and supervision with dedicated funding, test environments and regional hubs so clinicians, researchers and startups can pilot AI tools under real‑world oversight instead of guessing at compliance as they scale up.

Level / Region	Key initiative or funding
Federal	National AI & data strategy; BIPT as AI Act supervisor; sandboxes/living labs (Coalition Agreement)
Flanders	AI action plan ~EUR 32M/year (EUR15M companies, EUR12M research, EUR5M support); AI Research Programme (EUR12M/year)
Wallonia	DigitalWallonia4.ai (~EUR18M/year); ARIAC/TRAIL research project (EUR32M, 2021–2026)
Brussels	Innoviris support (~EUR22M dedicated budget; ~EUR44M invested since 2017) and targeted R&D calls

Belgium sits in the middle of a fast‑moving EU compliance story: the national picture for the EU AI Act is still classified as “unclear/awaiting designation,” with the Federal Public Service Economy publishing a list of 26 bodies and an Ethics Advisory Council on Data and AI already appointed, so health tech teams should be tracking formal names and remits as they emerge via the national implementation tracker (EU AI Act national implementation plans overview).

Important hard dates make the choices concrete - Member States must designate market surveillance and notifying authorities by 2 August 2025, and every state must host at least one AI regulatory sandbox by 2 August 2026 - meaning hospitals, medtech startups and vendors have a short runway to map who will supervise “high‑risk” clinical AI and where safe pilots can legally run (EU AI regulatory sandbox approaches and member state overview).

For Belgian healthcare this matters because sandboxes and named supervisors are the practical routes to test diagnostic algorithms or telemonitoring tools under regulatory guidance (and temporary protection from administrative fines if guidance is followed), so the looming deadlines are less theoretical than a compliance countdown that health teams can use to plan pilots, documentation and governance.

Item	Belgium (status / note)
National competent authorities	Awaiting designation / status: Unclear
Published bodies	List of 26 bodies published by Federal Public Service Economy
Ethics oversight	Ethics Advisory Council on Data & AI appointed
Key EU deadlines	Designate authorities by 2 Aug 2025; establish sandboxes by 2 Aug 2026

“The new law aims to foster the development and uptake of safe and trustworthy AI systems across the EU's single market.” - Council of the European Union press release on the AI Act

Health data in Belgium sits at the strict end of the GDPR spectrum: processing is treated as a “special category” and is generally prohibited unless one of the Article 9 exceptions applies (explicit consent, healthcare provision, substantial public interest, research, etc.) - a legal framework that teams must map early in any AI pilot (Article 9 GDPR special categories and exceptions).

National practice follows this tight line: Belgium did not create a broad national exception and has layered complementary measures - controllers must document who may access genetic, biometric or health data and bind those roles to confidentiality obligations - so clinical AI projects can't rely on ambiguity when reusing records for model training (Belgium health data rules and safeguards for clinical AI).

At the same time, national programmes (BIHR) and the incoming European Health Data Space aim to make safe secondary use for research and public‑health analytics possible, but only with purpose‑limitation, technical safeguards (pseudonymisation/federated approaches) and documented DPIAs - in short, secondary use is feasible, not free, and compliance failures carry real risk (regulators can impose administrative fines up to the GDPR caps and corrective measures).

The practical takeaway for Belgian teams: design governance and access lists before moving data; prototype with federated or de‑identified sets; and treat the BIHR/EHDS rollout as the compliance pathway that unlocks ethical, scalable secondary use (Belgium digital health laws, BIHR and EHDS guidance).

Topic	Key point
Special‑category health data	Generally prohibited; only lawful under Art.9 exceptions (consent, healthcare, public interest, research).
Belgian safeguards	No broad national exemption; controllers must list access categories and ensure confidentiality; DPA enforces GDPR.
Secondary use / BIHR & EHDS	Permitted for research/public health with Art.89 safeguards, DPIAs, pseudonymisation or federated methods; BIHR/EHDS are the governance path to scale.

The heart of Belgium's interoperability push is the interfederal eHealth Plan (2025–2027) and the Belgian Integrated Health Record (BIHR), a true “Moonshot” to stitch hundreds of systems into a virtual patient record so clinicians - and patients - can see the same medication lists, summaries and eventually images across settings; the BIHR aims for full national interoperability by 2026–2027 and is explicitly being aligned with the European Health Data Space after the EHDS Regulation entered into force on 26 March 2025, so national rollout and cross‑border technical work should be followed closely in the Digital Healthcare 2025 review.

While Belgium tops the EU eHealth Indicator Study (100%), that score masks practical limits - non‑standard inputs, fragmented portals and lagging imaging and implant coverage - which is why the program stresses standards (FHIR, SNOMED‑CT), federated approaches and change management rather than only plumbing; for concise legal and programme context see the EU eHealth coverage and the ICLG country practises on digital health.

Successful pilots will pair clear governance, the HeDERA/HDA pathways for safe secondary use, and concrete project management so interoperability becomes usable at the bedside, not just a policy line on a roadmap.

Item	Note
BIHR interoperability target	Full national interoperability by 2026–2027
EHDS	Regulation entered into force 26 March 2025; BIHR–EHDS interconnection planned from 2025
eHealth Plan	Interfederal eHealth Action Plan (NIHDI) for 2025–2027
EU eHealth Indicator	Belgium scored 100% (study reflecting status to 31 Dec 2024)

Protecting Belgian health services from cyberattack is now a legal and operational imperative: NIS2's Belgian transposition (Law of 26 April 2024, in force since 18 October 2024) pulls hospitals, larger labs and many digital health providers into an expanded scope that demands board‑level accountability, 24‑hour early warnings and a 72‑hour incident report cadence, stronger supply‑chain checks, multi‑factor authentication and tested continuity plans - in short, cyber risk becomes a clinical‑safety issue as much as an IT one.

Practical steps for Belgian healthcare teams include registering with the Centre for Cybersecurity Belgium (CCB), running an Article‑21 style gap analysis and hardening vendor contracts; several trackers note registration windows in March 2025 and audits starting later in 2025, so readiness can't wait (see a country implementation summary on CyberUpgrade and the practical deadlines flagged by Allegiance Law).

“a successful cyberattack costs a company on average €200,000, and 60% of SMEs go bankrupt within six months of a serious breach”

So pairing NIS2 controls with clinical risk governance is the fastest route to keeping AI‑enabled care safe and resilient (detailed requirements and timelines are usefully summarised by DataGuard and Cresco Cybersecurity).

Item	Key Belgium points
Transposition	Law adopted 26 Apr 2024; in force 18 Oct 2024 (Belgian NIS2 law)
Registration	Register with CCB (reported deadlines in March 2025: 18–31 Mar 2025)
Audits / enforcement	Formal compliance checks / audits from Q3 2025 (registration verification 2026–2027)
Core obligations	Risk management, incident reporting (24h/72h), supply‑chain security, MFA, business continuity
Sanctions	Up to €10M or 2% global turnover for essential entities

Reimbursement is the real gateway for AI to move from pilot to practice in Belgium: the National Institute for Health and Disability Insurance (NIHDI/RIZIV‑INAMI) runs a positive reimbursement list and demand that pricing and reimbursement dossiers are lodged together, so regulatory readiness is as important as clinical validation - see the detailed pricing & reimbursement framework in the Global Legal Insights review.

Belgium's pragmatic mHealth pyramid gives a defined path for digital therapeutics and AI apps (Level 3 can win provisional reimbursement, typically a three‑year paid trial while outcomes are collected), see the approval guidance on the NIHDI/mHealth route; this is why policymakers and vendors talk about a

paid clinical internship

for software.

Pockets of success already exist - NIHDI's new telemonitoring framework covers around 30 institutions and eight Belgian telemonitoring apps have cleared reimbursement hurdles (examples include moveUP and FibriCheck) - yet many innovators still hit a reimbursement bottleneck.

The market outlook is encouraging but realistic: Digital Healthcare 2025 projects a Belgian e‑health market of €754.53M in 2025 rising to €991.91M by 2029 (7.09% CAGR), so teams that pair measurable clinical benefit, clear integration into care pathways and early NIHDI engagement stand the best chance of turning AI pilots into reimbursed, scalable services.

Item	Key point (Belgium)
Reimbursement authority	NIHDI (RIZIV/INAMI) - positive reimbursement list; pricing & reimbursement submitted together (Global Legal Insights)
mHealth Level 3 (provisional)	Provisional reimbursement usually 3 years with evidence collection and price negotiation (mHealth pyramid guidance)
Telemonitoring	New NIHDI framework: ~30 institutions with agreements; 8 Belgian apps approved (e.g., moveUP, FibriCheck)
Market outlook	€754.53M (2025) → €991.91M (2029), CAGR 7.09% (Digital Healthcare 2025)

Practical AI in Belgian care is already moving from pilots to bedside tools: imaging and oncology are the clearest entry points, where the Cancer Image Europe platform (deployed by the EUCAIM project) will let Belgian hospitals and researchers access federated cancer imaging cohorts - today the hub links 57 datasets and ~47,000 subjects with a roadmap to >100,000 cases by 2026 - to validate detection and screening models under EHDS‑aligned safeguards (Cancer Image Europe (EUCAIM federated cancer imaging platform)); meanwhile the June 2025 EU study on AI in healthcare crystallises practical wins for radiology and operations - think triage engines that flag suspected pneumothorax and shave minutes off critical reads, ambient scribe tools that cut documentation time, and predictive models that spot sepsis or readmission risk before crises hit (2025 European Commission AI in Healthcare final report for radiology and clinical operations).

For Belgian teams the playbook is concrete: start with narrowly scoped pilots tied to measurable clinical or workflow gains, validate models on local scanners and BIHR/EHDS‑compatible feeds, and use federated infrastructures and upcoming EU funding calls to scale - so AI becomes a reliable clinical assistant, not a one‑off demo.

Use case	Practical benefit	Supporting initiative
Cancer imaging & screening	Cross‑site validation with large, diverse image sets (faster, more precise diagnostics)	Cancer Image Europe / EUCAIM (DIGITAL programme)
Radiology workflow & triage	Prioritise urgent scans, reduce backlog and clinician administration time	EU AI study recommendations; commercial AI triage tools
Predictive analytics (sepsis, readmissions, staffing)	Early warnings, optimized staffing, fewer avoidable admissions	Predictive platforms and Horizon/Apply AI funding streams

“If we progressively align this type of projects like EUCAIM or EUCanScreen with the HealthData@EU infrastructure of the European health data space, then we can bring cancer data at the forefront and make their reuse one of the first success stories of the European health data space.” - Lorena Boix Alonso

Viewed against European peers, Belgium is a striking mix: strong research muscle and R&D spend give it a top‑10 place in healthcare innovation, yet real‑world AI adoption still needs a push - only about 12% of Belgian employees say AI already affects their work, so hospitals and vendors must bridge the skills gap before pilots scale into routine care.

Belgium's #9 ranking in the FREOPP World Index of Healthcare Innovation highlights that research strength and cancer survival outcomes give a strong foundation (FREOPP 2022 World Index of Healthcare Innovation - Belgium #9), while other EU countries offer contrasting market‑access models to learn from - Germany's DiGA prescription‑and‑reimbursement pathway, the UK's NICE evidence standards and France's broad reimbursement of remote consultations show how policy design can turn pilots into paid services (Sidley Austin: Top Trends in the European Digital Health and AI Market).

At the same time the continental market is expanding rapidly (Europe's AI‑in‑healthcare market was valued in the low billions in 2024), so Belgium's comparative advantage is clear: convert world‑class labs and generous R&D into clinician‑facing, reimbursable tools by pairing pilots with NIHDI pathways, workforce upskilling and the interoperability work already under way (European AI in Healthcare Market Forecast - MarketDataForecast).

The vivid reality: strong science alone won't change bedside workflows - policy design, payment routes and training are the levers that will.

Metric	Value / note
FREOPP WIHI rank (Belgium)	#9 - score 51.99
Europe AI in healthcare market (2024)	USD 7.92 billion (MarketDataForecast)
Projected GenAI economic boost (Belgium)	Up to €50 billion over the next decade (Implement Consulting Group)

“Generative AI will boost global economic growth in the coming decade. It can increase productivity and boost Belgium's competitiveness. To capture the next wave of AI benefits across society, Belgium needs to promote innovation, invest in skills and ensure clear rules.” - Martin Thelle, Implement Consulting Group

Practical next steps for beginners in Belgian healthcare boil down to three simple moves: take inventory, follow a checklist, and build skills - start by cataloguing any AI touching patients or BIHR/EHDS feeds, classify each system's risk (many clinical tools will be “high‑risk”), and run a gap analysis so documentation, DPIAs/FRIAs and conformity tracks are ready before key deadlines (note: the AI Act brings heavy penalties - up to €35M or 7% of global turnover - so this isn't optional).

Use an EU AI Act checklist to convert legal obligations into a short, tactical roadmap (for example, the Vanta checklist and industry guides break down registration, QMS and post‑market monitoring), pair governance with cybersecurity and NIS2 readiness, and train staff so human oversight and audit trails aren't afterthoughts.

Start pilots narrowly (local scanners, federated testing, BIHR‑compatible datasets), log everything for audits, and invest in practical upskilling - courses like the AI Essentials for Work bootcamp syllabus teach non‑technical clinicians and managers how to use AI tools responsibly and write effective prompts while preparing teams for compliance and NIHDI engagement; that combination - clear paperwork, small measurable pilots, and real workplace training - turns regulatory risk into a launchpad for safe, reimbursable AI at the bedside.

Bootcamp	Length	Early bird cost	Registration
AI Essentials for Work bootcamp syllabus	15 Weeks	$3,582	Register for the AI Essentials for Work bootcamp