A.I. Governance in Belgium

In the preparation of my PromptParlour meetings, I wanted to get an idea of who in Belgian government is responsible for understanding and forecasting the (exponential) progress of Artificial Intelligence, and its impact on Belgian business, innovation and society. Who do Belgian/Flemish ministers call when they want an expert opinion on legislation or government intervention?

This report was done by Google Gemini Deep Research. Emphasis and annotations by me.

Executive Summary

Belgium has engineered a sophisticated and multi-layered governance architecture for Artificial Intelligence (AI), characterized by a strategic division of labor between its federal and regional governments. This structure operates as a “dual-engine” model: the Flemish Region serves as a dynamic, well-funded engine for research, innovation, and economic implementation, while the Federal Government provides the national chassis for coordination, legal harmonization, and regulatory oversight, largely driven by European Union mandates. This report provides a comprehensive analysis of the key people, institutions, and strategic frameworks responsible for researching AI, monitoring its evolution, and advising government bodies on its implications for legislation and public initiatives.

Federal (Belgium)

At the federal level, the Federal Public Service (FPS) Policy and Support (BOSA) (FOD Beleid en Ondersteuning) is the central coordinating body. Through its five core missions—including an AI Observatory, strategic policy support, and management of the AI4Belgium coalition—BOSA orchestrates the national dialogue and ensures alignment with the National Convergence Plan for AI/Nationaal convergentieplan voor de ontwikkeling van artificiële intelligentie. This plan, a high-level strategic document, aims to create a “#SmartAINation” by focusing on broad objectives such as trustworthy AI, cybersecurity, and economic competitiveness. The federal government has also proactively established a regulatory framework in anticipation of the EU AI Act, designating the Belgian Institute for Postal Services and Telecommunications (BIPT) as the lead supervisory authority and creating dedicated advisory committees for ethics and strategy.

Regional (Flanders)

In contrast, the Government of Flanders has implemented a highly structured and generously funded Flanders AI Policy Plan, committing approximately €36 million annually to a clear, three-pillar strategy.

• Pillar I focuses on strategic research through the Flanders AI Research Program, a powerful consortium of the region’s five universities and six strategic research centers, including the world-renowned imec.
• Pillar II drives implementation, with Flanders Innovation & Entrepreneurship (VLAIO) stimulating adoption in the private sector through subsidies and advice, and the AI Expertise Center within Digitaal Vlaanderen promoting AI use within the public administration.
• Pillar III provides the crucial societal framework, led by the Knowledge Centre Data & Society/Kenniscentrum Data & Maatschappij, which offers expert guidance on the ethical, legal, and social dimensions of AI.

The ecosystem is further strengthened by key institutions like imec, which provides the foundational hardware and nanoelectronics research essential for future AI advancements, and leading academic hubs such as KU Leuven (Leuven.AI). Policy advice flows through a multi-vector system of formal steering groups, expert committees, influential coalitions, and citizen engagement panels. The overarching influence of the EU AI Act serves as a powerful forcing function, compelling a degree of national cohesion and collaboration between the federal and Flemish levels. Ultimately, Belgium’s approach represents a compelling model for governing a transformative technology within a complex federal state, balancing regional economic ambition with national regulatory responsibility to position itself as a key player in Europe’s trustworthy AI landscape.

The Full Report

1. The Federal AI Strategy and Governance Framework

The Belgian federal government’s approach to Artificial Intelligence is primarily one of strategic orchestration, coordination, and regulatory harmonization. Rather than directly funding large-scale research and development, the federal apparatus focuses on creating a cohesive national framework that aligns the efforts of its diverse regions and communities, ensures compliance with overarching European Union legislation, and positions Belgium as a unified actor on the international stage. This strategy is embodied in the National Convergence Plan for AI and executed through a network of public services, coalitions, and newly established advisory and supervisory bodies.

1.1. The National Convergence Plan for AI: Forging a #SmartAINation

The cornerstone of the federal AI strategy is the National Convergence Plan for the Development of Artificial Intelligence, a comprehensive document approved by the Council of Ministers on October 28, 2022.1 The plan’s central ambition is to transform Belgium into a “#SmartAINation” by systematically harnessing the opportunities presented by AI across various sectors of the economy and society. The very title of the plan is significant; the term “convergence” underscores its primary function, which is to align and unify the disparate AI initiatives emerging from Belgium’s complex, multi-level governance structure, ensuring that federal, regional, and community efforts move in a coherent direction.

The plan’s vision is articulated through nine core objectives, which collectively reveal the federal government’s strategic priorities. These pillars are broad and cross-cutting, reflecting a high-level coordinating role rather than the direct, project-focused funding characteristic of the regional strategies, particularly that of Flanders. The nine objectives are

1. Promoting Trustworthy AI: Ensuring that AI is developed and used in a responsible, human-centric manner that respects fundamental rights and ethical principles.
2. Guaranteeing Cybersecurity: Addressing the new security risks and vulnerabilities introduced by the widespread adoption of AI and connected devices.
3. Boosting Belgium’s Competitiveness and Attractiveness through AI: Strengthening the national economy by fostering an environment conducive to AI innovation and investment.
4. Developing a Data-Driven Economy and a High-Performance Infrastructure: Building the foundational data and computational infrastructure necessary for a thriving AI ecosystem.
5. AI at the Heart of Healthcare: Leveraging AI to improve diagnostics, treatment, and the overall efficiency of the healthcare system.
6. Driving More Sustainable Mobility: Using AI to optimize transportation systems, reduce congestion, and enhance safety.
7. Protecting the Environment: Applying AI to address climate change and promote a circular economy.
8. Better Lifelong Training: Preparing the workforce for the labor market transformations driven by AI through continuous education and skills development.
9. Providing Citizens with Better Services and Protection: Using AI to improve the efficiency, accessibility, and personalization of public services.

To oversee the execution of this ambitious plan, a dedicated steering committee has been established. This committee is jointly managed by two key federal bodies: the Federal Public Service (FPS) Policy and Support (BOSA) and the FPS Economy. It is designed to function as the central point of contact at the federal level for all matters related to AI, tasked with formulating proposals on the AI program’s operation and ensuring its strategic implementation.1 The committee is chaired by the Director-General of the Directorate-General (DG) Simplification and Digitisation within FPS BOSA, placing the digital transformation arm of the federal government at the heart of national AI governance.

1.2. FPS BOSA: The Federal Engine for Digital Transformation

The Federal Public Service Policy and Support (BOSA) stands out as the primary institutional driver of the federal government’s digital and AI agenda. Its mandate is not limited to passive implementation; BOSA is tasked with actively shaping the digital evolution of the federal administration by monitoring technological trends, providing high-level strategic support to policymakers, and fostering an environment of innovation. This central role is operationalized through five core missions specifically focused on AI and emerging technologies:

1. AI and Digital Technologies Observatory: This mission serves as the federal government’s primary intelligence-gathering function. The observatory proactively monitors and analyzes emerging digital trends to anticipate their evolution and inform public strategy. It actively tracks AI and blockchain projects across Belgium, developing tools such as initiative maps that provide a comprehensive overview of the national ecosystem. This function directly addresses the need to “follow up on AI evolution and breakthroughs”.
2. Ethical and Legal Support: BOSA provides expert guidance on the complex ethical and legal dimensions of AI and blockchain technologies. This support is offered at all levels of governance—regional, national, and international—to ensure that the deployment of these technologies complies with established standards and regulations. This mission is central to advising the government on “safety and directions to take for legislation.”
3. Strategic Support for Political Decision-Makers (AI4Gov): This function establishes a direct and formal channel of communication between technical expertise and political leadership. Through the AI4Gov initiative, BOSA provides strategic advice and support directly to policymakers on issues related to AI, ensuring that political decisions are well-informed and strategically sound.
4. Artificial Intelligence Expertise Center and AI Lab: BOSA actively fosters innovation by developing cutting-edge AI solutions and conducting in-depth research. This center serves as an internal incubator for the federal government, developing and piloting AI applications such as chatbots for public services and tools for auditing the accessibility of government websites.
5. Ecosystem Management and International Outreach: BOSA plays a crucial role as a national convener. It leads and facilitates the AI4Belgium coalition, fostering collaboration among stakeholders from Flanders, Wallonia, and Brussels. Furthermore, it coordinates the steering committee responsible for implementing the National Convergence Plan, cementing its position as the institutional heart of federal-regional cooperation on AI.

The key entity within BOSA responsible for these tasks is the Directorate-General for Simplification and Digitalisation (DG SD), also known as the DG Digital Transformation. (…), the most recent information from BOSA’s official website indicates that Liliane Verreyen is currently the acting Director-General a.i. Political oversight is provided by Minister Vanessa Matz, who is in charge of Digitalisation, and Secretary of State Mathieu Michel, who is responsible for Digitalisation and Administrative Simplification.

1.3. AI4Belgium: A National Coalition for Stakeholder Alignment

AI4Belgium is a cornerstone of the federal strategy, operating as a “grassroot” community-driven coalition managed and led by FPS BOSA. It is not a formal government agency but rather a dynamic and influential ecosystem manager designed to connect and align the diverse stakeholders in the Belgian AI landscape, including academia, industry, and government bodies from all three regions.

The coalition’s primary functions are to co-animate the Belgian AI community, facilitate access to EU funding, propose concrete actions for AI training, and promote the adoption of AI technologies by industry. Its influence on formal policy is substantial; the “AI4Belgium roadmap” served as a foundational document that directly informed the development of the National Convergence Plan, demonstrating how this community-based initiative provides critical input into the official policymaking process. With a modest annual budget of less than €1 million, AI4Belgium’s value lies not in direct funding but in its ability to build consensus, share best practices, and serve as the primary informal channel for advice and strategic alignment across Belgium’s federated structure.

1.4. Regulatory and Ethical Oversight at the Federal Level

In response to the growing complexity of AI and the impending enforcement of comprehensive EU regulations, the federal government has moved to establish a clear and robust architecture for regulatory and ethical oversight. This structure is designed to ensure compliance, provide expert guidance, and create a predictable legal environment for AI innovation.

The most significant development is the designation of the Belgian Institute for Postal Services and Telecommunications (BIPT) as the lead national supervisory authority for the EU AI Act and the Data Act.12 This appointment, outlined in the 2025-2029 federal government agreement, represents a major expansion of BIPT’s traditional mandate as a regulator of the telecommunications sector. It positions the BIPT as the central enforcer of AI regulations in Belgium, responsible for market surveillance and ensuring that AI systems placed on the market comply with EU law.

To complement this supervisory function, the government has instituted two distinct consultative bodies to provide structured, expert advice

1. Federal Government Data Ethics and Artificial Intelligence Advisory Committee: Officially appointed by ministerial decree on May 8, 2024, this committee is housed within FPS BOSA.14 Its mandate is to provide scientific advice on the ethical, legal, economic, social, and environmental issues related to the use of AI within the federal administration. Its focus is therefore on internal governance, ensuring that the government itself is a responsible user of AI technology.
2. Artificial Intelligence Steering Committee: Established by a Royal Decree on May 24, 2024, this body has a broader mandate to advise the federal government on all other matters relating to artificial intelligence.14 It functions as a high-level strategic advisory council, providing guidance on national AI policy, technological developments, and economic opportunities.

The establishment of these distinct bodies—a single powerful regulator (BIPT) and two specialized advisory committees—demonstrates a deliberate and proactive effort to construct a comprehensive national governance framework. This architecture is clearly designed in anticipation of the full implementation of the EU AI Act, which requires Member States to designate competent authorities and establish clear governance structures.11 This indicates that European regulation is a primary driver of federal AI policy, with the federal government positioning itself as the key national entity for implementing and enforcing pan-European digital law.

2. The Flemish AI Ecosystem: A Three-Pillar Approach

While the federal government provides the national framework, the Flemish Region has established itself as the primary engine of AI research, development, and economic implementation in Belgium. The Government of Flanders has crafted a highly structured, ambitious, and exceptionally well-funded strategy designed to position the region as a European leader in the field. This strategy is not merely a policy document but an operational blueprint that creates a clear and integrated pipeline from fundamental research to industrial application and societal oversight.

2.1. The Flanders AI Policy Plan: A Regional Blueprint for Leadership

The central instrument of the Flemish strategy is the Flanders Artificial Intelligence (AI) Policy Plan. First launched in July 2019 and renewed for a second five-year cycle in March 2024, the plan’s explicit goal is to expand the region’s AI knowledge base, cultivate deep expertise within Flemish industry, and actively foster the widespread rollout of AI technologies to ensure Flanders assumes a “leading role” in the global AI revolution.

This ambition is backed by a substantial and sustained financial commitment. The plan represents an annual investment of approximately €32-36 million, a figure that dwarfs the federal government’s direct spending on AI initiatives and signals the high strategic priority placed on AI by the Flemish government. Political leadership for the plan rests at the highest level of the Flemish government. The responsible minister is the Minister-President of Flanders, a role that combines the portfolios of Economy, Innovation, and Industry. The current Minister-President is Matthias Diependaele. The primary administrative body overseeing the plan is the Department of Economy, Science and Innovation (EWI).

The defining feature of the Flanders AI Policy Plan is its clear and logical three-pillar structure. This architecture deliberately divides responsibilities and funding into three complementary and interconnected sections, creating a comprehensive ecosystem that addresses the full lifecycle of technological innovation

1. A research section, focused on strengthening strategic basic research.
2. An implementation and digitalization section, aimed at encouraging the adoption of AI by businesses and public services.
3. An accompanying section, which provides the necessary ethical framework, training, and public awareness initiatives.

This tripartite structure is the operational heart of the Flemish strategy, providing a clear roadmap from scientific discovery to market application and responsible governance.

2.2. Pillar I - Strategic Research: The Flanders AI Research Program

The first pillar of the Flemish plan is dedicated to fundamental scientific advancement and is embodied by the Flanders AI Research Program. This program is the core research engine of the regional strategy, receiving an annual budget of €14.185 million to conduct strategic basic research into generic AI methodologies. Its research is intentionally demand-driven, designed to produce foundational knowledge that can be applied across a wide range of applications in healthcare, industry, and government.

The program’s research agenda is organized around five key focus themes, selected to align with both global trends and Flanders’ existing research strengths:

• Responsible AI: Developing AI that is fair, transparent, explainable, and aligned with societal values.
• Human-centered AI: Creating AI systems that support human decision-making and interact safely and intuitively with users.
• Sustainable AI: Designing energy-efficient AI hardware and software to minimize environmental impact.
• Productive and Data-efficient AI: Building AI systems that can learn from small datasets and are easily integrated into existing workflows.
• Resilient and Performant AI: Creating dynamic and adaptive AI that can operate effectively in uncertain and changing environments.

The immense strength of the Flanders AI Research Program lies in its collaborative structure. It is not a single institution but a powerful consortium of Flanders’ eleven leading research institutions. This brings together the region’s five universities—KU Leuven, Ghent University, University of Antwerp, Hasselt University, and Vrije Universiteit Brussel—with its six strategic research centers: imec, Flanders Make, VIB (Flanders Institute for Biotechnology), VITO (Flemish Institute for Technological Research), Sirris (the collective center of the Belgian technology industry), and ILVO (Flanders Research Institute for Agriculture, Fisheries and Food). This “triple helix” model, uniting over 300 researchers, ensures that cutting-edge academic research is closely linked with applied science and industrial needs from the very beginning.

2.3. Pillar II - Implementation & Adoption: The Role of VLAIO and Digitaal Vlaanderen

The second pillar of the Flemish plan is focused on ensuring that the knowledge generated by the research program is effectively translated into tangible economic and societal value. This is achieved through a dual-pronged approach, with dedicated agencies targeting the private and public sectors, respectively. This separation allows for tailored strategies that recognize the different needs, incentives, and constraints of businesses versus government bodies.

Flanders Innovation & Entrepreneurship (VLAIO) is the primary vehicle for driving AI adoption in the private sector. As the Flemish government’s central point of contact for entrepreneurs, VLAIO is tasked with de-risking and accelerating the integration of AI into Flemish businesses. It receives approximately €9.5 million annually from the AI plan, supplemented by a separate €12 million fund for broader SME digitalization, to carry out its mission. VLAIO’s activities include raising awareness, providing expert advice, and offering direct financial support through a range of subsidy instruments for research, development, and implementation projects. A specific funding stream, known as “VAIOP,” is earmarked for projects that actively disseminate the results of the Flanders AI Research Program to the broader business community, creating a direct, funded pathway from basic research to industrial application.

A key figure representing VLAIO’s role in the broader ecosystem is Jeroen Fiers, an advisor who sits on both the VAIA steering committee and the overarching AI Steering Group for the Flemish plan.

The internal counterpart to VLAIO is the AI Expertise Center, housed within Digitaal Vlaanderen, the agency responsible for the digital transformation of the Flemish public sector.42 The center’s mission is to promote and guide the responsible adoption of AI within the Flemish government and local authorities. Its work is structured around five pillars:

1. Vision & Strategy, which involves developing frameworks like the AI Playbook;
2. Innovation, which includes piloting new tools such as Microsoft 365 Copilot for public servants;
3. Reliable AI, providing crucial guidance on compliance with the EU AI Act;
4. People & Organization, focusing on change management and AI literacy for government employees; and 5; AI Architecture, developing scalable and reusable technical solutions.

The center also manages the AI Adviesraad (AI Advisory Council), a strategic body composed of experts from academia, industry, and government that advises the Flemish government on AI policy and implementation. A key individual supporting the center’s strategic work is Hans Arents, a senior advisor at Digitaal Vlaanderen.

2.4. Pillar III - Societal Framework: Ethics, Training, and Awareness

The third and final pillar of the Flemish plan provides the essential societal guardrails and enablers for the first two. With an annual budget of €5 million, it focuses on building a trustworthy, skilled, and engaged society capable of navigating the complexities of the AI revolution. This is accomplished through three distinct but complementary initiatives.

The Knowledge Centre Data & Society (Kenniscentrum Data & Maatschappij) serves as the ethical and legal conscience of the Flemish AI ecosystem. It is a unique collaboration between three leading university research groups: imec-SMIT at Vrije Universiteit Brussel (VUB), imec-MICT at Ghent University, and the Centre for IT & IP Law (CiTiP) at KU Leuven. Its core mission is to provide concrete, actionable guidance on the legal, ethical, and social implications of AI. It does this by developing practical tools (such as its “EU AI Act in 6 steps” guide and the “ALLY” framework for responsible AI strategy), publishing accessible research, and offering advice to policymakers, companies, and the public to ensure that AI is implemented in a socially responsible and legally compliant manner. The center is led by Academic Coordinator Dr. Rob Heyman and Operational Coordinator Pieter Duysburgh.

The Flemish AI Academy (VAIA) addresses the critical need for skills and training. As a collaboration between all Flemish universities and university colleges, VAIA’s mission is to provide a comprehensive and accessible range of AI courses and training programs for professionals across all sectors. It acts as a central hub for lifelong learning in AI, ensuring that the Flemish workforce is equipped with the necessary skills to adapt to and thrive in an AI-driven economy.

Finally, Amai! is a citizen-facing initiative designed to foster public awareness and engagement. It informs citizens about AI and actively involves them in the co-creation of AI solutions to pressing societal challenges in areas like climate, mobility, and health. This ensures that the development of AI in Flanders is not a purely top-down process but is instead informed by public needs and values.

3. Key Institutions and Personalities in AI Research and Policy

The strategic frameworks established by the federal and Flemish governments are brought to life by a network of world-class research institutions and a cadre of influential leaders. These organizations and individuals form the intellectual and operational core of the Belgian AI ecosystem, driving the research, executing the policy, and bridging the gaps between science, industry, and governance.

3.1. Leading Research Institutions

At the heart of Belgium’s AI capabilities are several research institutions with global reputations. Their work provides the foundational knowledge and technological breakthroughs that fuel the entire ecosystem.

• imec (Interuniversity Microelectronics Centre) is arguably the most critical technological institution in the Belgian AI landscape. Headquartered in Leuven, imec is a world-leading independent R&D hub in nanoelectronics and digital technologies. Although it was founded with support from the Flemish Government, it operates as a truly international organization with a global partner network. Imec’s role in AI is foundational and strategic; it focuses on developing the next-generation hardware—the smaller, faster, more powerful, and more energy-efficient microchips—that is essential for enabling future AI applications.56 This focus on the physical layer of AI distinguishes the Belgian ecosystem, as it is not merely concerned with applying existing AI software but with building the fundamental technology for the next generation of AI. Imec is a key partner in the Flanders AI Research Program and also coordinates the Flanders AI EDIH (European Digital Innovation Hub), which helps accelerate AI adoption among SMEs.

The leadership at imec is crucial to this vision. The President and CEO is Luc Van den hove, a pivotal figure who has shaped imec’s global ecosystem, championed its role in the EU Chips Act, and consistently emphasized the need to future-proof AI through hardware innovation.55 Within his leadership team, two individuals are central to the AI strategy: Karl Tuyls, who serves as an imec fellow and the Chief Technology Officer (CTO) for AI, and Steven Latré, who holds the position of AI Strategy Advisor.

University Hubs form the backbone of fundamental AI research in the region.

• KU Leuven is a powerhouse in AI research and is recognized as one of Europe’s top universities for spin-off activities. Its Leuven.AI institute serves as a central hub, coordinating multidisciplinary AI research across dozens of labs and research groups. As a leading member of the Flanders AI Research Program, KU Leuven is central to the region’s scientific output. A key personality is Prof. Bart De Moor, a full professor at KU Leuven who is not only a promotor for the Flanders AI Research Program but also serves as the influential chair of the VAIA (Flemish AI Academy) steering committee, creating a strong link between top-tier research and professional training.
• The VUB AI Lab at Vrije Universiteit Brussel is renowned for its pioneering and interdisciplinary research, with a long history in areas such as autonomous systems, reinforcement learning, and AI ethics. It plays a leading role within the Flanders AI Research Program, heading the research challenge on “Multi-agent collaborative AI”.
• Ghent University, the University of Antwerp, and Hasselt University are the other vital academic partners in the Flanders AI Research Program

Each brings specialized expertise, and their collective participation ensures a comprehensive and robust research base that covers the full spectrum of AI disciplines.

3.2. Governmental Leadership and Key Figures

The direction and implementation of AI policy are guided by key political and administrative leaders at both the federal and Flemish levels.

At the Federal Level:

• Ministerial Leadership: The political direction for the federal digital agenda is set by Minister Vanessa Matz, who is the Minister of the Modernisation of the Public Administration and is in charge of Digitalisation, and Secretary of State Mathieu Michel, who is responsible for Digitalisation and Administrative Simplification.
• Administrative Leadership: The day-to-day execution of the federal AI strategy falls to FPS BOSA, where Liliane Verreyen serves as the acting Director-General for Simplification and Digitalisation

At the Flemish Level:

• Ministerial Leadership: The ultimate political authority for the comprehensive and well-funded Flanders AI Policy Plan is Matthias Diependaele. In his dual role as Minister-President of the Government of Flanders and Flemish Minister for Economy, Innovation, Industry, and Digitalisation, he oversees the entire AI ecosystem in the region.
• Administrative Leadership: Several key figures are instrumental in managing the Flemish ecosystem. Jan Smedts, as the head of Digitaal Vlaanderen, chairs the AI Advisory Council that guides public sector adoption. Jeroen Fiers, an advisor at VLAIO, is a crucial link, representing the implementation pillar in the overarching AI Steering Group.26Dr. Rob Heyman, the Academic Coordinator of the Knowledge Centre Data & Society, is a highly influential voice on the ethical and legal dimensions of AI and also participates as an expert in the AI Steering Group.26The Belgian AI ecosystem is held together not just by formal structures but by a network of these key individuals who often hold multiple roles across different institutions. For instance, Prof. Bart De Moor’s leadership in both the research program and the AI academy ensures that scientific advancements are quickly translated into training curricula. Similarly, the presence of figures like Jeroen Fiers and Rob Heyman on the central AI Steering Group guarantees that the perspectives of industrial implementation and ethical oversight are tightly integrated into the highest level of strategic decision-making. This overlapping leadership model fosters strong informal and formal links, reducing institutional silos and enabling rapid, coherent action across the different pillars of the Flemish plan.

4. The Flow of Advice: From Research to Legislation

The process by which research findings, ethical considerations, and stakeholder feedback are translated into government policy and legislation in Belgium is a multi-vector system. It involves a sophisticated interplay of formal advisory bodies, influential informal coalitions, direct academic input, and structured citizen engagement. This complex network ensures that decision-making on AI is informed by a wide range of perspectives, balancing expert opinion with industry needs, ethical principles, and public sentiment.

4.1. Formal Pathways of Influence

Both the federal and Flemish governments have established formal structures to channel expert advice directly into the policymaking process.In the Flemish Government, the primary formal advisory body is the AI Steering Group. This group is responsible for the overall management of the Flanders AI Policy Plan and is composed of high-level representatives from universities, key industry associations (Agoria, Voka, UNIZO), research centers, and government agencies, including VLAIO and Digitaal Vlaanderen.26 Advice flows vertically from the three pillars of the AI plan into this central committee. The Flanders AI Research Program provides scientific and technological insights; VLAIO channels feedback on the needs and challenges of industrial implementation; and the Knowledge Centre Data & Society delivers crucial analysis on ethical and legal issues. The steering group synthesizes this input and provides direct advice to the Department of Economy, Science and Innovation (EWI) and the responsible minister, Matthias Diependaele. For matters specific to the public sector, Digitaal Vlaanderen’s AI Advisory Council offers targeted recommendations on AI adoption within the government itself.

At the Federal Government level, the advisory process has been recently formalized and strengthened. The “Strategic support for political decision-makers” (AI4Gov) mission within FPS BOSA represents a direct, ongoing channel of expert advice to government ministers.2 This has been supplemented by the creation of two new, highly structured committees. The Federal Data and AI Ethics Advisory Committee provides specific, scientific advice on the responsible use of AI within federal government operations, focusing on internal governance. In contrast, the broader Federal AI Steering Committee is tasked with advising the government on all other matters related to AI, serving as the main source of high-level strategic guidance on national policy.

4.2. Inter-Governmental Collaboration and EU Influence

Given Belgium’s federal structure, mechanisms for inter-governmental collaboration are essential. The primary vehicle for this is the AI4Belgium coalition, managed by FPS BOSA. It serves as the main forum for dialogue and consensus-building between the federal government and the regions (Flanders, Wallonia, and Brussels), ensuring that regional initiatives are understood at the national level and vice versa. The National Convergence Plan itself is the intended outcome of this collaborative ideal, aiming to create a unified national vision.However, the most powerful driver of national cohesion is the external pressure of European Union legislation. The EU AI Act, a comprehensive and legally binding regulation, acts as a significant forcing function for federal-regional alignment.11 The federal government is responsible for the national implementation of the Act, including designating the supervisory authority (BIPT) and ensuring a consistent enforcement framework across the country.12 Simultaneously, regional bodies are actively preparing their stakeholders for compliance. The AI Expertise Center within Digitaal Vlaanderen is developing guidance specifically for the Flemish public administration on how to navigate the Act’s requirements. The Knowledge Centre Data & Society plays a pivotal role in interpreting the complex legal text and developing practical tools to help Flemish companies and organizations understand their obligations.12 This shared, non-negotiable requirement for compliance necessitates a common understanding and a coordinated response, compelling a level of collaboration that might otherwise be challenging to achieve.

4.3. Citizen and Stakeholder Engagement

Both levels of government have also recognized the importance of incorporating public perspectives into AI governance, moving beyond purely expert-driven advice.At the federal level, the FPS Foreign Affairs organized a pioneering citizens’ panel in the lead-up to the finalization of the EU AI Act.65 A group of 60 Belgians, randomly selected to be representative of the national population in terms of age, gender, language, and other demographics, was convened to deliberate on the promises and risks of AI. Their deliberations produced nine key messages for policymakers, emphasizing the need for permanent human oversight in decision-making, greater consideration for the environmental impact of AI, and the importance of ensuring AI serves the common good without undermining democracy or employment. These recommendations were intended to inform the Belgian government’s negotiating positions and to inspire the strategic agenda of the European Commission.65In Flanders, citizen engagement is designed as a continuous process through the Amai! initiative, which is part of the third pillar of the AI Policy Plan.18 Amai! actively involves the public in the co-creation of AI solutions for specific societal problems, such as those in mobility or healthcare. This approach ensures that public needs and concerns are not just solicited as feedback on existing policies but are integrated into the innovation process from the very beginning, helping to build public trust and ensure that AI development is aligned with societal values.

5. Analysis and Conclusion

Belgium’s approach to the governance of Artificial Intelligence is a direct reflection of its unique federal structure. It is a complex, multi-layered system that, rather than being a source of fragmentation, has been strategically leveraged to create a “dual-engine” model for innovation and oversight. This architecture combines the strengths of regional agility and economic focus with the stability of a national regulatory framework, positioning the country to navigate the challenges and opportunities of the AI era.

5.1. Strengths: The “Dual-Engine” Advantage

The primary strength of the Belgian system is its effective division of labor, which allows it to function as a dual-engine entity. The Flemish Region acts as a powerful and well-funded R&D and economic engine. Through its highly structured, three-pillar AI Policy Plan, Flanders has created an integrated pipeline that efficiently moves ideas from fundamental research in its world-class universities and research centers, through industrial implementation facilitated by VLAIO, all while being guided by the ethical oversight of the Knowledge Centre Data & Society. This model is designed to maximize innovation, economic growth, and competitiveness.Simultaneously, the Federal Government provides the essential chassis and steering mechanism for the entire nation. Its role is not to compete with the regions in funding innovation but to provide the overarching legal framework, national and international representation, and regulatory stability required to guide the system. By focusing on “convergence,” leading the AI4Belgium coalition, and establishing the national supervisory architecture for the EU AI Act, the federal level ensures legal coherence, prevents a fragmented internal market, and presents a unified face to Europe and the world. This symbiotic relationship allows for both dynamic, regionally-tailored innovation and consistent, nationwide regulatory standards.

5.2. Challenges: Navigating Complexity and Ensuring Convergence

The principal challenge of this model is its inherent complexity. The multiplicity of plans, agencies, advisory bodies, and research programs across different levels of government can be difficult to navigate, particularly for external stakeholders such as international investors, SMEs, or researchers seeking to engage with the ecosystem. The success of the National Convergence Plan will depend entirely on the continued effectiveness of coordinating bodies like the joint federal steering committee and the AI4Belgium coalition. Without robust and continuous communication and alignment, there remains a risk of policy divergence, duplication of effort, and inefficient allocation of resources.Furthermore, the significant expansion of BIPT’s mandate from a telecommunications regulator to the national supervisor for the AI Act will be a critical test. The institute will need to rapidly build deep technical and legal expertise in AI governance to effectively carry out its new responsibilities. Ensuring that its enforcement actions are consistent and well-coordinated with the innovation-focused activities of the regions will be crucial to maintaining a balance between regulation and growth.