🇧🇪Belgium

Overview

Belgium's approach to Artificial Intelligence (AI) regulation is multifaceted, reflecting its complex federal structure where both federal and regional governments play significant roles in shaping the digital landscape. The country does not currently have a single, dedicated national legislative act specifically governing AI, but rather relies on a combination of strategic policy documents, research programs, and collaborative initiatives to guide AI development and adoption. This decentralized yet coordinated strategy aims to position Belgium as a 'SmartAI Nation' that prioritizes trustworthy, human-centric AI. The overall maturity level of AI governance in Belgium is rapidly evolving, largely driven by the direct applicability of the European Union's AI Act, which entered into force in August 2024 with various provisions taking effect incrementally. Belgian federal and regional entities are actively working to align their national strategies and prepare stakeholders for the compliance requirements introduced by this landmark EU legislation. Key initiatives, such as the AI4Belgium coalition and regional AI policy plans in Flanders and Wallonia, focus on fostering a robust AI ecosystem through investment in research, industrial adoption, ethical oversight, and skills development. These efforts underscore a commitment to leveraging AI for economic growth and societal benefit while upholding fundamental rights and ethical principles.

Regulatory approach

Belgium's regulatory approach to AI is primarily characterized by a blend of soft law instruments at the national and regional levels, complemented by the directly binding nature of the EU AI Act. At the federal level, strategic documents like the National Convergence Plan for the Development of Artificial Intelligence (SmartAI.Nation) and the AI4Belgium roadmap set out cross-cutting priorities and foster collaboration among government, industry, academia, and civil society. These initiatives are non-coercive, defining objectives, funding windows, and monitoring mechanisms rather than prescriptive legal rules. Their enforcement relies on collaborative governance, voluntary adoption of best practices, and targeted public funding to incentivize responsible AI development and deployment. The forthcoming full implementation of the EU AI Act introduces a horizontal, risk-based regulatory framework that will directly apply across all EU member states, including Belgium. This Act bans certain harmful AI practices and imposes strict obligations on high-risk and general-purpose AI systems, moving beyond soft law to binding legal requirements. Belgian regional policies, such as the Flanders AI Policy Plan and DigitalWallonia4.ai, are designed to bridge fundamental research with industrial application while ensuring alignment with the EU AI Act. They emphasize ethical considerations, data governance, and social well-being, often through the development of guidelines, assessment tools (like ALTAI-based assessments), and capacity-building programs to prepare local stakeholders for the evolving regulatory landscape. Belgium's AI governance framework is characterized by a multi-layered structure involving federal and regional entities, reflecting the country's division of powers. At the federal level, the FPS Policy and Support (BOSA) plays a central coordinating role, driving the digital evolution and reforms within the federal government. BOSA manages the AI4Belgium coalition, which acts as a central ecosystem manager connecting academia, business, and the public sector, and contributes to the monitoring of the National Convergence Plan for AI. Within BOSA, a Federal Government Data Ethics and Artificial Intelligence Advisory Committee was officially appointed in May 2024 to provide scientific advice on the ethical, legal, economic, social, and environmental issues related to AI use within the federal administration. Additionally, an Artificial Intelligence Steering Committee was established by Royal Decree in May 2024 to advise the federal government on broader AI matters, including national AI policy and technological developments. With the advent of the EU AI Act, the Belgian Institute for Postal Services and Telecommunications (BIPT/IBPT) has been designated as the lead national market surveillance authority, responsible for ensuring that AI systems comply with EU law. The Belgian Data Protection Authority (APD/GBA) is the primary body for enforcing the GDPR and the national data protection act, and it also publishes guidelines on AI, particularly concerning privacy implications. At the regional level, the Flanders Department of Economy, Science and Innovation (EWI) oversees the Flanders AI Policy Plan and the Flemish AI Research Programme, working with VLAIO (Flanders Innovation & Entrepreneurship) to support businesses and with the Knowledge Center Data & Society for ethical and legal oversight. In Wallonia, the Agence du Numérique (AdN) coordinates the DigitalWallonia4.ai program, while in the Brussels-Capital Region, Innoviris and Paradigm.brussels co-coordinate the FARI AI for the Common Good Institute, which focuses on multidisciplinary AI research for urban challenges.

Enforcement & penalties

The landscape of penalties and enforcement for AI in Belgium is primarily shaped by the General Data Protection Regulation (GDPR) and, increasingly, by the European Union AI Act. For violations of the GDPR, which are often triggered by AI processing of personal data, administrative fines can reach up to €20 million or 4% of a company's total worldwide annual turnover, whichever is higher. The Belgian Data Protection Authority (APD/GBA) is responsible for enforcing these provisions at the national level. With the EU AI Act entering into force, a new tiered penalty system is being introduced, comparable to the GDPR but with potentially higher maximum administrative fines. Non-compliance with prohibited AI practices, such as social scoring or real-time biometric identification (with narrow exceptions), can result in administrative fines of up to €35 million or 7% of global annual turnover, whichever is higher. Non-compliance with specific provisions related to high-risk AI systems or providers of general-purpose AI can lead to fines of up to €15 million or 3% of worldwide annual turnover. Providing incorrect, incomplete, or misleading information to authorities may incur fines of up to €7.5 million or 1% of worldwide annual turnover. Member states are required to establish national rules on penalties and enforcement measures by August 2, 2025, and Belgium is currently in the process of discussing its supervisory structure for the AI Act. For the various national and regional AI policy plans and roadmaps, enforcement mechanisms are typically non-coercive, relying on contractual obligations for grant recipients, reputational incentives, and alignment with broader EU standards, rather than direct statutory sanctions.

Data protection

Belgium's data protection framework is robust, primarily rooted in the European Union's General Data Protection Regulation (GDPR), which is directly applicable across all member states. The GDPR establishes comprehensive rules for the processing of personal data, granting individuals significant rights over their data and imposing strict obligations on data controllers and processors. This includes principles such as lawfulness, fairness, transparency, purpose limitation, data minimization, accuracy, storage limitation, integrity, confidentiality, and accountability. AI systems operating in Belgium that process personal data must fully comply with these GDPR requirements, including provisions related to automated decision-making and profiling. Complementing the GDPR, the Belgian Data Protection Act of 2018 further details and adapts certain GDPR provisions to the national context, establishing the Belgian Data Protection Authority (APD/GBA) as the independent supervisory authority. The APD/GBA is tasked with enforcing data protection laws, handling complaints, and providing guidance, including specific guidelines on AI and its impact on privacy. Regional AI policies and initiatives, such as the Flanders AI Policy Plan and the FARI AI for the Common Good Institute, explicitly integrate data protection principles like 'privacy by design' and robust data governance into their frameworks, ensuring that AI development and deployment respect fundamental rights and societal norms.

Sector-specific rules

While Belgium does not currently possess dedicated national legislative acts for AI that are specific to individual sectors, the various federal and regional AI policy initiatives actively promote and pilot AI applications within key economic and societal domains. These strategic frameworks, such as the Flanders AI Policy Plan and DigitalWallonia4.ai, identify specific sectors for targeted support and innovation. For instance, areas like healthcare (e.g., improving diagnostic accuracy, optimizing hospital resource management, personalized care), Industry 4.0 (e.g., industrial automation), smart logistics, urban planning (e.g., 'Digital Twin' projects for traffic flow optimization), mobility, and public services (e.g., administrative simplification, employment services, animal welfare monitoring) are central to the implementation strategies of these regional programs. In these sectors, the application of AI is guided by the overarching principles of trustworthy AI, ethical considerations, and data protection as defined by the GDPR and the forthcoming EU AI Act. While specific AI legislation for these sectors is absent, existing sectoral regulations (e.g., health regulations, product safety laws) would apply to AI systems deployed within them. For example, general rules on liability for defective products would apply to faulty AI systems in the absence of specific AI liability laws. The policy documents encourage the development of sector-specific use cases and sandboxes to test governance and safety mechanisms, ensuring that AI tools are tailored to meet actual market needs and societal requirements while adhering to ethical and legal standards.

International alignment

Belgium demonstrates strong international alignment in its approach to AI regulation, primarily through its deep integration within the European Union framework. The country is directly subject to the EU AI Act, which establishes harmonized rules for AI across all member states, and its national and regional strategies are explicitly designed to prepare for and complement this landmark legislation. This alignment ensures that Belgium's AI ecosystem operates within a globally recognized and comprehensive regulatory environment for trustworthy AI. The EU AI Act's risk-based approach, banning of certain harmful practices, and obligations for high-risk AI systems directly influence Belgian policy development and implementation. Beyond the EU AI Act, Belgium also aligns its AI legal and policy framework with other key international standards. The OECD AI Principles, which advocate for human-centric, transparent, and accountable AI, are explicitly endorsed and cited as a guiding framework for Belgium's National Convergence Plan for AI (SmartAI.Nation). This commitment to OECD principles ensures that Belgium's national strategy aligns with international best practices for responsible AI governance. Furthermore, Belgium actively participates in broader EU initiatives such as Horizon Europe, Digital Europe, European Digital Innovation Hubs (EDIHs), and AI Factories, leveraging regional funding to secure additional resources and expertise from the wider EU ecosystem. This integrated approach ensures that Belgium remains competitive and attractive for AI innovation on a global scale, contributing to and benefiting from the collective European effort in digital transformation.

What's next

The future of AI regulation in Belgium will be significantly shaped by the ongoing implementation and full application of the European Union AI Act. While some core provisions, such as the ban on certain unacceptable AI practices and AI literacy obligations, came into effect in February 2025, the majority of the Act, including rules on high-risk AI systems, will apply from August 2, 2026. This phased implementation means that Belgian stakeholders, including businesses and public administrations, will continue to adapt their AI development and deployment practices to ensure full compliance with the new binding legal requirements. A critical upcoming development is the designation of national competent authorities and the establishment of specific rules on penalties and fines by member states, which must be in place by August 2, 2025. Belgium is currently in the process of discussing and formalizing its supervisory structure to effectively enforce the AI Act. At the federal level, the recently established Federal Government Data Ethics and Artificial Intelligence Advisory Committee (May 2024) and the Artificial Intelligence Steering Committee (May 2024) will continue to provide scientific and strategic advice to the government, influencing future policy directions and ensuring ethical considerations remain at the forefront of AI deployment within the federal administration. Regional AI initiatives, such as the Flanders AI Policy Plan 2024-2028 and the Wallonia DigitalWallonia4.ai Program, will continue their strategic efforts in fostering research, innovation, and ethical adoption, evolving their programs to align with and complement the EU AI Act. These ongoing developments underscore a proactive and adaptive regulatory environment in Belgium, aiming to navigate the complexities of AI while maximizing its benefits for society and the economy.