European Union - AI Policy Approach

Overview

The Communication "Artificial Intelligence for Europe" (COM(2018)237 final), published on 25 April 2018 by the European Commission, sets out a first comprehensive EU policy approach to AI. It frames AI as a strategic enabling technology and proposes a three‑pillar approach: step up investment (public and private), prepare for socio‑economic changes, and ensure an ethical and legal framework that builds trust. The Communication is explicitly a policy and coordination instrument rather than binding legislation: it launches consultative mechanisms such as the European AI Alliance and the High‑Level Expert Group on AI to develop ethics guidance, and it signals the intent to prepare a Coordinated Plan with Member States and other stakeholders. The document links AI policy to funding and infrastructure initiatives, including research programmes and initiatives on data and computing resources, and identifies priority areas for deployment and skills. It also provided the basis for subsequent EU outputs including the Coordinated Plan on AI (Dec 2018), ethics guidelines developed by the AI HLEG, and later policy instruments. The full official text is available on the EU legal and publications portals, including EUR-Lex (COM(2018)237).

Definitions

The Communication adopts an operational understanding of artificial intelligence as systems that display intelligent behaviour by analysing their environment and taking actions — with some degree of autonomy — to achieve specific goals. AI systems can be purely software‑based (e.g., voice assistants, image analysis, search engines) or embedded in hardware devices (e.g., robots, autonomous vehicles, drones, IoT applications). The Communication distinguishes between development (research, algorithms, models, computing capacity) and deployment (applications in sectors such as healthcare, transport, manufacturing and public services), while recognising cross‑cutting issues such as data, standards, safety, privacy and ethics. While the Communication does not create a formal legal definition to be used in binding legislation, it established vocabulary and scope used in subsequent EU instruments.

Governance and Institutional Framework

The Communication proposes a multi‑layered governance and institutional approach combining EU institutions, Member States, expert groups and multistakeholder platforms. At EU level the European Commission (especially DG CONNECT) coordinates actions and funding, and the Communication immediately triggered the establishment of the European AI Alliance and the Commission's High‑Level Expert Group on Artificial Intelligence (AI HLEG) to provide guidance on ethics and policy. The Communication calls for a Coordinated Plan between the Commission and Member States to pool resources, align national strategies and accelerate deployment, and it recommends that Member States set national strategies and designate contact points for coordination. It also envisions support mechanisms including Digital Innovation Hubs, testing and experimentation facilities, and networks for research infrastructures. The document emphasises the role of sectoral regulators and standardisation bodies to update or adapt existing sectoral safety rules and conformity mechanisms. The approach is cooperative: the Communication positions the Commission as coordinator and facilitator rather than the origin of binding rules, while signalling that gaps not covered by existing law may be addressed by future legislative proposals.

Key Focus Areas

The Communication identifies several priority areas where EU action should focus. First, research, technology and innovation: increasing investment in AI research, providing access to high‑performance computing and data, and supporting startups and SMEs to commercialise AI. Second, skills and labour: preparing the workforce for changing job requirements through reskilling and upskilling and supporting education. Third, data and infrastructure: improving data sharing and access while respecting privacy and data protection obligations, and developing secure computing infrastructure. Fourth, ethics, safety and trust: ensuring AI respects fundamental rights, non‑discrimination, transparency and accountability; the Communication calls for ethical guidance and tools to assess trustworthiness and safety. Fifth, public sector uptake: encouraging Member States and public administrations to adopt AI in public services to improve efficiency and the quality of services. Sixth, standards, testing and experimentation: fostering common standards, testbeds, sandboxes and certification schemes to build interoperable, safe and reliable systems. Finally, international cooperation: coordinating with global partners on research, standards and governance. These focus areas are cross‑cutting: investments, governance, legal frameworks and societal considerations are interdependent and designed to promote "AI that works for people." The Communication articulated concrete domains — healthcare, transport, energy, manufacturing and public sector services — where EU advantage could be built by combining excellence and trust.

Implementation Framework

To implement its objectives the Communication proposes a combination of soft law coordination, funding instruments and targeted initiatives. It commits the Commission to prepare a Coordinated Plan with Member States to align national strategies and investments, leverage EU research programmes (notably Horizon 2020 at that time) and digital infrastructure, and mobilise private investment. The Communication recommends establishing or strengthening national contact points and Digital Innovation Hubs to support SMEs and public authorities. It also envisages deploying testing and experimentation facilities, promoting open data initiatives while protecting personal data, and supporting standards development via European standardisation bodies. Implementation relies on staged collaboration: the Commission provides leadership, technical support and funding channels, the AI HLEG and Alliance deliver ethics guidance and stakeholder input, and Member States translate strategic objectives into national measures and investments. The Communication expressly encourages Member States to increase public procurement of AI solutions and to facilitate demonstration projects that create markets and build trust.

Monitoring and Evaluation

The Communication foresees monitoring through the Coordinated Plan, reporting by Member States, and feedback loops via the European AI Alliance and expert groups. It calls for indicators to track investment, research outputs, skills deployment and uptake in the public and private sectors. The Commission committed to follow‑up on progress, to update the plan as necessary and to publish information on investments and activities. Moreover, the Communication envisaged the use of policy experimentation, pilot projects and standardisation timelines that can be evaluated against predefined objectives. While the Communication itself does not create binding reporting obligations, it establishes the governance channels through which monitoring, assessment and iterative adjustments would take place and feed into subsequent initiatives and legislative proposals.

Penalties, Liability, and Appeals

As a policy Communication, COM(2018)237 is non‑binding and does not itself establish penalties, administrative sanctions or judicial appeals mechanisms. Instead, it identifies liability, safety and legal certainty as areas requiring further analysis and potential legislative or regulatory action. The Communication laid the groundwork for subsequent work on liability and adaptation of existing sectoral safety rules; later EU proposals and legislative acts (for example, the EU AI regulatory package proposed in subsequent years) address enforcement, conformity assessment and penalties. Therefore, penalties and appeals would be set out in later binding instruments rather than in the Communication itself.

Relationship to Other Instruments

The Communication sits at the start of a sequence of EU AI policy outputs. It explicitly anticipated a Coordinated Plan (delivered in December 2018), the formation of the AI HLEG and the development of ethics guidelines, and it paved the way for the 2020–2021 White Paper and the subsequent regulatory proposals that addressed legal frameworks, liability and sectoral safety rules. The Communication references existing EU legal instruments (e.g., data protection under the GDPR) and calls for complementary action where gaps exist. It also links AI objectives to EU funding instruments and digital policy priorities and envisioned interactions with sectoral rules (transport safety, medical devices, product safety and liability frameworks), standardisation bodies and international fora. Official follow‑ups include the Coordinated Plan on AI (COM(2018)795) and the AI HLEG ethics guidelines.

International Alignment

The Communication recognises that global cooperation is essential for research, standards and ethical norms. It encourages international collaboration with likeminded partners to develop common approaches to governance, share best practices for research and testing, and converge on standards that facilitate interoperability and trade. The Communication also references multilateral initiatives and the need to engage in global debates on AI ethics, safety and trade. It positions the EU as seeking to promote a human‑centric, trustworthy AI model grounded in European values, while cooperating internationally to avoid fragmentation and to reinforce Europe’s competitiveness. The Commission indicated engagement with G7/G20 dialogues, international standardisation organisations and transatlantic partnerships to align principles and technical standards.

Implementation Timeline

Sources and References