European Union - AI and Human Rights (2024/2218)

Overview

Council Decision (EU) 2024/2218 of 28 August 2024 authorises the signing, on behalf of the European Union, of the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law. The Decision is brief and procedural: Article 1 grants authorisation to sign the Convention subject to its final conclusion; Article 2 tasks the European Commission with ensuring the signature on behalf of the Union; and Article 3 states that the Decision enters into force on the date of adoption (28 August 2024). The Decision was published in the Official Journal (OJ L 2024/2218, 4.9.2024) and complements prior Council authorisation to open negotiations (Council Decision (EU) 2022/2349). The authorised signature took place when the Commission signed the Convention on behalf of the EU on 5 September 2024 during the Conference of Council of Europe Ministers of Justice in Vilnius; see the European Commission statement and Council of Europe information pages for contextual materials and the adopted text of the Convention. For the official Decision text see Council Decision (EU) 2024/2218 - EUR-Lex and for the Convention details see the Council of Europe portal at The Framework Convention on Artificial Intelligence - Council of Europe.

Definitions

For purposes of this Decision and the connected Convention the principal definitions are procedural and conform to EU treaty language. "Signing" means the formal act by which the European Union, as a regional organisation, expresses its consent to be bound by the Convention subject to subsequent conclusion or internal consent procedures. "Convention" or "Framework Convention" refers to the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (adopted by the Council of Europe Committee of Ministers on 17 May 2024 and opened for signature on 5 September 2024). "Commission" refers to the European Commission acting on behalf of the Union under Article 218(5) TFEU. "Conclusion" denotes the later act by which the Union gives its consent to be bound under the internal procedures required by the Treaties. The Decision uses the term "subject to its conclusion" to indicate that signature does not itself constitute consent to be bound under international law without the Union's later internal approval processes.

Governance and Institutional Framework

The Decision sits within the Union's external relations legal architecture. Under Article 218 TFEU the Council normally authorises the opening of negotiations and adopts decisions authorising signature and conclusion of international agreements. The present Decision was adopted by the Council of the European Union after a Commission proposal; it builds on Council Decision (EU) 2022/2349 which authorised the opening of negotiations on behalf of the Union. Institutional roles are therefore: the Council (adopter of the Decision), the Commission (negotiator and signatory agent), and the European Parliament (consulted as required by Treaty provisions where the subject matter so requires). The Treaty legal basis cited is Article 114 in conjunction with Article 218(5) TFEU. Post-signature, the Commission is tasked by the Decision with ensuring the actual signature event and subsequent coordination. Implementation of any obligations resulting from the Convention (after the EU gives consent to be bound) will require inter-institutional coordination between the Commission, the Council and Member States and close alignment with the EU's internal AI regulatory framework, notably the AI Act (Regulation (EU) 2024/1689). For Commission statements about signature see the EU External Action Service press release at EEAS: European Commission signs historic Framework Convention on AI, and for the Council of Europe background see Committee on Artificial Intelligence (CAI) - Council of Europe.

Key Focus Areas

While the Council Decision itself is procedural, the Convention it authorises the EU to sign addresses multiple substantive areas aligned with EU priorities. Key focus areas include protection of fundamental rights (freedom of expression, privacy, non-discrimination), democratic safeguards (protection against manipulative uses of AI in political processes), rule of law guarantees (transparency, possibility of judicial review and effective remedies), risk-based approaches to AI governance, data governance consistent with data protection rules, documentation and accountability standards (e.g., model cards, logs), transparency obligations for AI-generated content and human oversight requirements. The Convention seeks to be technology-neutral and lifecycle-focused: obligations and principles are designed to apply across development, deployment and use of systems rather than being limited to particular techniques. The Convention also emphasises international cooperation, exchange of best practices, capacity building and support for regulatory sandboxes and innovation-friendly measures. These themes mirror and complement provisions in the EU AI Act and existing human-rights law frameworks; the Convention is thus designed to create a common baseline of obligations for Parties while allowing domestic implementation adapted to national legal systems. For the Convention text and explanatory materials see the Council of Europe Convention page at Council of Europe - Framework Convention on AI.

Implementation Framework

The Decision's mandate to the Commission is limited to ensuring signature; detailed implementation responsibilities arise only if and when the Convention is concluded and the Union gives consent to be bound. At that stage, internal implementation will likely require: review of competence and division of powers between the EU and Member States; identification of measures requiring Union-level regulation (e.g., cross-border data flows, market-wide standards); alignment with existing EU instruments (notably the AI Act, GDPR and sectoral rules); and possible delegated or implementing acts to reconcile any technical standards. The Commission will be the primary coordinator at EU level, working with Member States through the Council preparatory bodies and working groups, and with the Council of Europe bodies responsible for monitoring the Convention (including the Committee on Artificial Intelligence). The Convention also envisions reporting, peer review and co-operation mechanisms for Parties which will interact with EU monitoring obligations. For the adopted EU AI Act see Regulation (EU) 2024/1689 (AI Act) - ELI, an instrument with which any Convention obligations must be reconciled.

Monitoring and Evaluation

The Decision itself does not create monitoring mechanisms; those are established by the Convention and by existing Council of Europe instruments. Subject to the EU's eventual consent to be bound, monitoring will likely involve reporting obligations on Parties, exchanges of information, and oversight by a Council of Europe monitoring body (or committee) specifically designated under the Convention. Within the EU, the Commission and Member States would need to coordinate reporting to any Council of Europe monitoring mechanisms and would evaluate domestic compliance through the frameworks already in place under the AI Act and other sectoral regimes. The Convention complements EU-level supervisory architecture by creating an international layer of peer review, technical assistance, and possible follow-up recommendations; the EU's internal monitoring (e.g., market surveillance, notified bodies, competent authorities under the AI Act) would remain the primary means of ensuring compliance in the Union's internal market.

Penalties, Liability, and Appeals

Council Decision (EU) 2024/2218 does not impose penalties or create liability regimes; it only authorises signature. The Convention (and subsequent implementing EU or national law) addresses remedies, liability and potential sanctions for breaches of Convention obligations. Enforcement within the EU will therefore depend on the interplay between the Convention's requirements and existing and future EU/national enforcement mechanisms (including administrative fines, civil liability remedies, and sector-specific sanctions). Any binding penalties for actors operating in the Union would derive from the AI Act, data protection law (GDPR), and national implementation measures. The Commission and Member States will need to identify whether additional civil or administrative liability provisions are required to give effect to specific Convention obligations post-ratification/consent.

Relationship to Other Instruments

The Decision references and must be read alongside prior EU actions: Council Decision (EU) 2022/2349 (authorising negotiations) and the EU AI Act (Regulation (EU) 2024/1689). The Convention is explicitly intended to complement existing international human-rights instruments (e.g., the European Convention on Human Rights) and EU law. Where the EU signs and later consents to be bound by the Convention, the Union will be responsible for ensuring that EU law is compatible with Convention obligations. Implementation will therefore require inter-instrument coordination to avoid conflicts, ensure legal certainty, and respect the division of competences between the Union and its Member States. Key references include: the Official Journal publication of the Decision (EUR-Lex), the Convention text and explanatory report on the Council of Europe website (CoE Convention page) and the EU AI Act (ELI - AI Act).

International Alignment

The Decision facilitates the EU's engagement with a multinational legally binding instrument on AI. The Convention's adoption and EU signature promote alignment between the Union's regulatory approach and an international standard-setting process led by the Council of Europe, which includes a broad set of states and observers. The Convention aims to harmonise baseline obligations for Parties and to serve as a reference for third countries and international fora. EU participation signals support for a human-rights centred, rule-of-law based global approach to AI governance and reinforces consistency between the EU AI Act and emerging international standards, while leaving open the necessary work to translate Convention provisions into EU and national law.

Implementation Timeline

Event	Date
Council of Europe adoption of Convention	2024-05-17
Council Decision (EU) adoption authorising signature	2024-08-28
Decision enters into force (date of adoption)	2024-08-28
Opening for signature (Vilnius Conference)	2024-09-05
European Commission signature on behalf of the EU	2024-09-05
Convention entry into force (subject to ratifications)	Upon required ratifications (per Convention: typically after X ratifications — consult Council of Europe instruments)

Compliance Checklist

Action	Responsible	Completed (Y/N)
Authorise EU signature	Council of the EU	Y (2024-08-28)
Ensure Commission signs on behalf of EU	European Commission	Y (2024-09-05)
Assess compatibility with AI Act and Union law	Commission / Legal Services	N/ongoing
Coordinate Member State implementation planning	Council/Commission	N/ongoing
Prepare reporting and monitoring arrangements	Commission & Member States	N/ongoing
Identify gaps requiring legislative measures	Commission	N/ongoing

Sources and References

Source	Type
Council Decision (EU) 2024/2218 of 28 August 2024 (Official Journal / EUR-Lex)	Primary Source
The Framework Convention on Artificial Intelligence - Council of Europe	Primary Source
EEAS: European Commission signs historic Framework Convention on AI (press release)	Primary Source