Belgium - Flemish AI Research Programme

Overview

The Vlaams Onderzoeksprogramma Artificiële Intelligentie (AI Vlaanderen) is a Flemish strategic basic research programme that centralises top-level AI research across Flanders to advance foundational methods and accelerate uptake in industry and government. The programme was approved by the Flemish Government on 28 June 2019 and initially funded via the Department of Economy, Science and Innovation (EWI). Its implementation is coordinated with the innovation agency VLAIO and executed through a consortium of Flemish universities and research centres; these arrangements and the programme scope are described on official programme and department pages (see EWI decision 28 June 2019, VLAIO guidance and the programme web portal at Flanders AI Research). The programme targets four technological research challenges: data-efficient and actionable AI, energy-efficient/edge AI, multi-actor collaborative AI, and human-centred AI, and illustrates outcomes through demonstrators in domains such as health, industry (Industry 4.0) and public services.

Definitions

For the purposes of the programme and participating agreements, common definitions include: "AI research programme" as public-funded strategic basic research activities focused on algorithms, systems and demonstrators; "consortium partners" as accredited Flemish universities and recognised research centres collaborating under a joint plan; "demonstrator" as a research-built, domain-specific prototype or proof-of-concept used to validate methodologies; and "implementation instruments" referring to innovation uptake tools such as VLAIO's ICON mechanism. The programme operates within the legal and regulatory constraints of Belgian and EU law, including GDPR for personal data processing and applicable ethics/regulatory frameworks for health or safety subjects.

Governance and Institutional Framework

Governance is led by the Flemish Department EWI as policy owner and funder; EWI provides strategic oversight, budgetary decisions and policy alignment. Operational implementation and industry translation tasks are carried out by VLAIO through innovation instruments such as ICON, complemented by dedicated actors including the Vlaamse AI Academie (VAIA), the Data & Society knowledge centre and collaborative partners (see VLAIO and Flanders AI). Research execution and coordination are the responsibility of a multi-institutional consortium (five universities and multiple research centres such as imec, Flanders Make, VIB, VITO, Sirris). Programme governance uses grant agreements and consortium governance rules: project-level steering committees, reporting obligations, ethical review where needed and coordination with regional policy instruments. EWI retains budgetary and policy accountability, including formal decision records of the Flemish Government (e.g. the 28 June 2019 decision).

Key Focus Areas

The programme concentrates research effort on four core technological challenges: (1) AI-driven data science that supports hybrid, trusted and actionable decision-making using limited labelled data and domain knowledge; (2) edge and real-time AI that prioritises energy-efficiency and embedded intelligence for sensors and local devices; (3) multi-actor collaborative AI that enables autonomous and interoperable decision-making in ecosystems where no single actor has full system knowledge; and (4) human-centred AI that improves human–AI interaction, interpretability and trust. These technical pillars are paired with cross-cutting priorities: responsible and ethical AI (addressing privacy, bias and transparency), sustainability (energy- and resource-aware AI), and demonstrator-driven validation in application domains such as healthcare (e.g., disease progression models), industrial automation and public-sector chatbots and services. The programme funds both fundamental methods and exemplars (demonstrators) to create pathways for industrial adoption and social benefit. To facilitate uptake, the implementation strand uses instruments such as VLAIO’s ICON to bring research outputs to companies and practice, while parallel capacity-building measures (training, VAIA) develop human capital.

Implementation Framework

Implementation is organised across strategic research projects coordinated by the consortium with time-limited research grants and deliverable schedules. Grant contracts set requirements for work packages, milestone reporting, open-science obligations and demonstrator delivery; projects are expected to respect GDPR and research ethics procedures for human-subject or health-related research. The implementation framework also prescribes pathways for translation: VLAIO’s innovation instruments support business uptake and applied R&D, while the VAIA and knowledge centres provide training and dissemination. Project selection and funding allocation follow public grant rules and internal scientific evaluation; interview and review panels assess scientific merit, relevance to Flemish societal and economic priorities and potential for technology transfer. Partnership agreements govern IP management and exploitability of demonstrators to ensure that research results can be used by industry while preserving open-science goals where appropriate.

Monitoring and Evaluation

Monitoring relies on periodic reporting to EWI and VLAIO, milestone verification, publication records and demonstrator evaluations. Key performance indicators include publications in peer-reviewed outlets, demonstrator readiness levels, industry collaborations, patenting and training outputs (PhD/early-career researchers). The programme pursues iterative evaluation cycles to adapt focus areas, reallocate resources for emergent needs and measure societal impacts (e.g., technology diffusion in Flemish SMEs). EWI publishes decisions and news updates about programme milestones and funding decisions; VLAIO reports on ICON-facilitated uptake activities and supported companies. Independent or internal evaluations may be commissioned to assess effectiveness, alignment with policy objectives and the quality of knowledge transfer.

Penalties, Liability, and Appeals

The research programme is delivered via grant and contract law rather than regulatory sanctions. Contractual remedies for non-compliance typically include requirement to correct deficiencies, suspension of payments, termination of the grant agreement and clawback or repayment of monies where misuse is proven. Research integrity breaches, falsification or serious GDPR violations trigger contractual, institutional and potentially legal actions under Belgian law; projects with clinical or safety implications must comply with relevant sectoral regulations, and liability flows through existing institutional insurance and legal frameworks. Appeals against funding or administrative decisions are handled under administrative law procedures applicable to Flemish grant decisions; recipients may contest administrative decisions through the relevant departmental review or judicial channels under Belgian administrative procedure rules.

Relationship to Other Instruments

The programme acts as the research pillar of the broader Flemish AI policy and is linked to the general Vlaams Beleidsplan Artificiële Intelligentie (approved March 2019) and to implementation instruments such as VLAIO’s ICON programme for industry translation. It complements Flemish initiatives for capacity-building (Vlaamse AI Academie - VAIA), the Data & Society knowledge centre and international cooperation channels (NCP Flanders). Where research outputs touch regulated sectors (e.g., health, medical devices, energy or transport), project teams must also respect sector-specific regulatory regimes and certification pathways (e.g., medical device regulation, safety regimes). The programme is therefore deliberately designed as a research-and-innovation accelerator rather than a substitute for statutory regulation.

International Alignment

AI Vlaanderen explicitly seeks alignment with EU-level strategies and funding programmes to maximise interoperability and to position Flemish research within European research networks. The programme’s thematic choices (responsible AI, energy-efficiency, human-centred approaches) are consistent with EU priorities and Horizon Europe research themes. Consortium partners engage in international collaborations, co-funded projects and publications to ensure cross-border scientific exchange and to align research outputs with international standards and norms. Official programme communications note the role of the programme in supporting Flemish participation in EU initiatives and in adopting relevant recommendations from EU-level bodies.

Implementation Timeline

Event	Date	Notes
Vlaams AI policy plan approved (Flemish Government)	2019-03-22	High-level policy plan establishing three complementary pillars (policy, research, implementation)
Vlaams Onderzoeksprogramma AI Vlaanderen approved (Flemish Government)	2019-06-28	Subsidy decision for research programme (initial budgetary decision up to €6 million for 2019)
Programme operationalisation & consortium formation	2019-07 - 2020	Consortium of universities and research centres formed; projects initiated
Second-cycle programme (VAIOP) start	2024-01-01	Second research cycle (2024-2028) announced and running

Compliance Checklist

Requirement	Compliant Action
Grant agreement obligations	Maintain project plan, milestones, and periodic reporting to EWI/VLAIO
Research ethics	Obtain institutional review board / ethics approvals for human-subjects research
Data protection	Document GDPR compliance, DPIAs where required, secure processing and data minimisation
Open-science / publication	Publish research results per grant terms and deposit demonstrator documentation
IP & exploitation	Follow consortium IP arrangements, notify stakeholders of commercialization plans

Sources and References

Source	Type
Beslissingen Vlaamse Regering - DEWI (28 June 2019)	Primary Source
VLAIO - Thematic ICON & AI implementation	Primary Source
Flanders AI Research - programme portal	Primary Source
imec - Vlaams AI-Onderzoeksprogramma	Primary Source