How Albania, Bosnia and Herzegovina, Kosovo, Montenegro, North Macedonia and Serbia are aligning with the EU AI Act (Regulation (EU) 2024/1689), measured only against published official sources. Methodology v1.3. A country implementation score is published only once 80% of its weighted scored indicators carry complete, current evidence; below that the score is withheld rather than estimated.
Country evidence last checked Jul 16, 2026. Tracker data last updated Oct 3, 2026. 90 of 138 scored checks have a current research review, 17 with complete evidence and 73 where no official evidence was found. Drawn from 17 unique official source documents.
Country status
Albania
EU status: candidate. Lead institution: Council of Ministers and National Agency for Information Society. AI legislation: draft. National AI strategy: draft. Council of Europe Framework Convention on AI: signed Jun 15, 2026, not yet ratified.
Evidence coverage 4% of weighted scored indicators. 2 of 23 scored checks reviewed, 1 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
The draft AI law completed public consultation on 26 June 2026, and the draft National AI Strategy 2025-2030 and action plan completed consultation in September 2025. Adoption and entry into force were not evidenced as of 16 July 2026.
Bosnia and Herzegovina
EU status: candidate. Lead institution: Council of Ministers and competent entity institutions. AI legislation: not evidenced. National AI strategy: not evidenced. Council of Europe Framework Convention on AI: signed Dec 9, 2025, not yet ratified.
Evidence coverage 21% of weighted scored indicators. 23 of 23 scored checks reviewed, 20 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
Official 2025 reporting records no AI-specific law. Countrywide AI Act approximation, a national AI strategy, and an operational governance framework remain unevidenced. The Council of Europe AI Convention is signed but not recorded as ratified.
Kosovo
EU status: potential candidate. Lead institution: Office of the Prime Minister, Ministry of Digitalization and Public Administration, and Information and Privacy Agency. AI legislation: not evidenced. National AI strategy: not evidenced. Council of Europe Framework Convention on AI: not signed.
Evidence coverage 4% of weighted scored indicators. 1 of 23 scored checks reviewed, 0 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
An official AI Act readiness assessment published in June 2026 states that Kosovo has no dedicated AI law, no national AI strategy, and no formally designated AI competent authority. The assessment contains recommendations. Only the separate three-year IPA commitments are treated as formal future actions.
Montenegro
EU status: candidate. Lead institution: Ministry of Public Administration. AI legislation: planned. National AI strategy: draft. Council of Europe Framework Convention on AI: signed Nov 5, 2024, not yet ratified.
Evidence coverage 16% of weighted scored indicators. 21 of 23 scored checks reviewed, 19 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
A working group reviewed the first draft AI Strategy 2026-2030 in May 2026. European Commission reporting says AI Act approximation and the required enforcement structures still need to be established.
North Macedonia
EU status: candidate. Lead institution: Ministry of Digital Transformation. AI legislation: planned. National AI strategy: preparing. Council of Europe Framework Convention on AI: signed May 8, 2026, not yet ratified.
Evidence coverage 16% of weighted scored indicators. 21 of 23 scored checks reviewed, 19 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
SMART/MK 2030 includes an AI pillar and the Ministry of Digital Transformation reported a dedicated national AI strategy in preparation in May 2026. AI Act approximation and enforcement structures remain outstanding in EU reporting.
Serbia
EU status: candidate. Lead institution: Government of Serbia and Ministry of Science, Technological Development and Innovation. AI legislation: planned. National AI strategy: adopted. Council of Europe Framework Convention on AI: not signed.
Evidence coverage 36% of weighted scored indicators. 22 of 23 scored checks reviewed, 14 with no official evidence found. Implementation score withheld: evidence coverage is below the 80% publication threshold.
Serbia adopted its AI Strategy 2025-2030 in January 2025, including legal, institutional, literacy, high-risk impact, and public-sector support measures. European Commission reporting says AI Act approximation and operational governance structures are still required.
EU AI Act timeline
Aug 1, 2024 — European Union — EU AI Act entered into force (completed). The regulation entered into force across the European Union.
Jan 10, 2025 — Serbia — Serbia adopted AI Strategy 2025-2030 (completed). National AI strategy adopted.
Feb 2, 2025 — European Union — Prohibitions and AI literacy obligations applied (completed). Rules on prohibited practices and AI literacy began to apply.
Aug 2, 2025 — European Union — Governance and GPAI obligations applied (completed). Governance rules and general-purpose AI obligations began to apply.
Sep 11, 2025 — Albania — Albania draft AI Strategy consultation closed (completed). Consultation on the draft National AI Strategy 2025-2030 and action plan closed.
May 7, 2026 — European Union — Political agreement on the AI Omnibus (completed). Political agreement on the AI Omnibus simplification package and the revised high-risk implementation timetable. The Omnibus was subsequently adopted as Regulation (EU) 2026/1744 and entered into force on 27 July 2026.
May 27, 2026 — Montenegro — Montenegro reviewed first draft AI Strategy (completed). Working group discussed the first draft of the 2026-2030 strategy.
Jun 15, 2026 — Albania — Albania signed the Council of Europe AI Convention (completed). The signature is context for rights alignment and does not affect the EU AI Act score.
Jun 22, 2026 — Kosovo — Kosovo published an EU AI Act readiness assessment (completed). The official assessment records the baseline, explicit gaps, IPA commitments, and non-binding recommendations.
Jun 26, 2026 — Albania — Albania AI law consultation closed (completed). Public consultation on the draft law ran from 29 May to 26 June 2026.
Jul 9, 2026 — Albania — Albanian government sent AI Convention ratification bill to Parliament (completed). The government approved submission of the bill. Completed parliamentary ratification was not yet recorded.
Jul 20, 2026 — European Union — Commission guidelines on AI transparency obligations (Article 50) published (completed). The European Commission published guidelines on the transparency obligations of providers and deployers of AI systems under Article 50 of the AI Act. They give practical guidance on how the transparency rules apply. Commission guidelines are not legally binding and create no new obligation, so this date does not move any country score.
Jul 23, 2026 — European Union — Feedback closes on EU high-risk AI guidelines (completed). Targeted feedback on the Commission's draft high-risk classification guidelines closed. As of 2 August 2026 the final guidelines had not yet been adopted or published.
Jul 27, 2026 — European Union — AI Omnibus enters into force (completed). Regulation (EU) 2026/1744 (the AI Omnibus) entered into force. It adopts the simplification package and locks the phased high-risk timetable into binding law (Annex III standalone systems 2 December 2027 and product-embedded systems 2 August 2028). The 2 August 2026 application of most provisions and transparency rules is unchanged.
Aug 2, 2026 — European Union — General AI Act application date (completed). From 2 August 2026 the Commission's AI Office and national authorities begin enforcing the AI Act, and new transparency obligations start to apply. High-risk requirements follow the later phased timetable (2 December 2027 and 2 August 2028).
Dec 2, 2026 — European Union — Marking deadline for AI content systems already on the market (upcoming). New Article 111(4) of the AI Act: providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content that were placed on the market before 2 August 2026 must take the necessary steps to comply with the Article 50(2) machine-readable marking obligation by 2 December 2026. Systems placed on the market from 2 August 2026 are covered from that date.
Dec 2, 2026 — European Union — New AI Act prohibitions on intimate-image and CSAM generation apply (upcoming). Regulation (EU) 2026/1744 adds prohibited practices to Article 5 of the AI Act covering AI systems that generate or manipulate non-consensual intimate material or child sexual abuse material, with a safe harbour where the provider has reasonable and adequate preventive measures. The amended Article 113 applies Article 5(1), first subparagraph, points (ba) and (bb) and Article 5(1a) and (1b) from 2 December 2026.
Aug 2, 2027 — European Union — National AI regulatory sandboxes must be operational (upcoming). Article 57(1) of the AI Act as amended by Regulation (EU) 2026/1744 requires each Member State to ensure at least one national AI regulatory sandbox is operational by 2 August 2027, alone or jointly with other Member States. Western Balkans countries are not bound by this date, but it is the EU reference point for the tracker's sandbox and implementation-enabler indicators.
Aug 2, 2027 — European Union — Legacy general-purpose AI models must comply (upcoming). Article 111(3) of Regulation (EU) 2024/1689: providers of general-purpose AI models that were placed on the market before 2 August 2025 must take the necessary steps to comply with the obligations laid down in the Regulation by 2 August 2027. This closes the transition for models that predate the 2 August 2025 GPAI application date.
Dec 2, 2027 — European Union — Annex III high-risk rules (upcoming). Standalone Annex III high-risk obligations apply. Date set by the adopted AI Omnibus (Regulation (EU) 2026/1744).
Aug 2, 2028 — European Union — Product-integrated high-risk rules (upcoming). High-risk obligations for AI embedded in regulated products apply. Date set by the adopted AI Omnibus (Regulation (EU) 2026/1744).
How this is measured
Each country is scored across 23 scored indicators and 1 context-only indicator, grouped into 6 pillars totalling 100 points: Legal approximation (30 points); Governance and enforcement (20 points); High-risk AI controls (20 points); Transparency and GPAI (10 points); Data protection and rights (10 points); Implementation enablers (10 points). A check counts only with a published official source, an exact locator, an evidence note and a current verification date. Where the official corpus was searched and nothing was found, the check is recorded as "no official evidence found" — never as a zero. Coverage measures verified documentation, not implementation.
Published by AI Balkans under the Creative Commons Attribution 4.0 licence (https://creativecommons.org/licenses/by/4.0/). Reuse with attribution.