Latvia AI Regulation Overview

Overview

Latvia's approach to artificial intelligence regulation combines strategic policy roadmap planning with institutional capacity-building and targeted legal safeguards. Foundational directives were set out in the Cabinet-approved 2020 informative report 'On the Development of Artificial Intelligence Solutions' and expanded under the Digital Transformation Guidelines 2021–2027. Under the coordinating leadership of the Ministry of Smart Administration and Regional Development (VARAM), Latvia's policy focuses on accelerating trustworthy AI adoption across the economy and public service while protecting fundamental human rights, public security, and democratic integrity.

In March 2025, Latvia established a dedicated institutional pillar by enacting the Law on the Development of Artificial Intelligence and the Law on the Latvian Artificial Intelligence Centre (Mākslīgā intelekta centra likums). This framework created the Artificial Intelligence Centre (Mākslīgā intelekta centrs) as a national foundation designed to unify public, academic, and private sector efforts, manage AI tenders and dataset curation, and operate a controlled testing environment. Alongside supportive ecosystem measures, Latvia has enacted binding statutory restrictions, specifically criminalising election-focused deepfake manipulation and creating mandatory labeling and administrative removal regimes for AI-generated political content.

Regulatory Approach

Latvia follows a hybrid approach that pairs strategic, non-regulatory national coordination with targeted, binding legal interventions in high-risk domains. Rather than creating a standalone, all-encompassing technical AI code, the central statutory framework establishes the Artificial Intelligence Centre strictly as a coordinating, advisory, and enabling authority. Regulatory oversight remains distributed among existing sectoral ministries, statutory enforcement authorities, and administrative bodies.

To tackle concrete emerging harms, Latvia uses targeted legislative amendments to existing codes. High-risk technological interference in democratic processes is regulated directly through hard law: amendments to the Criminal Law address malicious synthetic media during election periods, while amendments to the Pre-election Agitation Law establish strict transparency obligations for political advertisers. This approach balances innovation support through regulatory sandboxing with enforceable safeguards against targeted societal risks.

Key AI Legislation

• Law on the Latvian Artificial Intelligence Centre (Mākslīgā intelekta centra likums, 2025): Establishes the Centre as a foundation governed by a nine-member Supervisory Board, authorises state funding and donations, sets operational rules, and creates a statutory special regulatory environment (regulatory sandbox) for controlled AI development and testing.
• Law on the Development of Artificial Intelligence (Mākslīgā intelekta attīstības likums, 2025): Defines national AI policy terms and legal parameters for national capacity mapping, advisory roles, and public-private coordination in alignment with EU standards.
• Amendments to the Criminal Law (Article 90.1, 2024): Criminalises the intentional creation or dissemination of false, discrediting deepfake audio-visual materials targeting political candidates or parties during pre-election campaign windows, imposing sentences of up to five years imprisonment.
• Amendments to the Pre-election Agitation Law (Grozījumi Priekšvēlēšanu aģitācijas likumā, 2024): Imposes mandatory clear labeling on paid political campaign materials containing AI-generated depictions of persons or synthetic events, prohibits automated anonymous account agitation, and grants KNAB removal authority.
• Digital Transformation Guidelines 2021–2027: National strategic policy defining key action directions, including digital skills, cybersecurity, connectivity, public service digital transformation, and ICT innovation hubs.
• Informative Report 'On the Development of Artificial Intelligence Solutions' (2020): Foundational policy report establishing Latvia's six-pillar AI approach and assigning VARAM as the primary coordinating ministry.

Governance & Enforcement Bodies

Institutional oversight for AI in Latvia involves both central strategic coordination and specialized law enforcement. The Ministry of Smart Administration and Regional Development (VARAM) acts as the primary policy coordinator for digital transformation, data governance, and strategic AI planning. Under VARAM's supervisory oversight, the Artificial Intelligence Centre operates as a dedicated foundation. The Centre is governed by a nine-member Supervisory Board representing VARAM, the Ministry of Economics, the Ministry of Defence, private industry, and higher education, with operational, legal, and IT security secretariat support supplied by the State Digital Development Agency.

Enforcement of mandatory statutory provisions is executed by existing regulatory authorities. The Corruption Prevention and Combating Bureau (KNAB) monitors political advertising compliance and holds direct administrative authority to prohibit and order the removal of non-compliant AI-generated campaign materials. Criminal investigations concerning malicious deepfakes and automated interference with elections fall under the jurisdiction of the State Police (Valsts policija) and the Prosecutor's Office (Ģenerālprokuratūra). Supervisory oversight regarding personal data processing across AI applications is exercised by the Data State Inspectorate (Datu valsts inspekcija, DVI).

Penalties & Enforcement

Penalties for AI-related violations are established under specialized criminal and administrative legal codes. Under Article 90.1 of the Criminal Law, deliberate manufacture or dissemination of false, discrediting deepfakes during pre-election periods carries severe criminal sanctions, including imprisonment for up to five years, temporary deprivation of liberty, probation supervision, or community service. The law also penalises the aggravated use of automated data processing systems to obstruct citizen participation in elections, referendums, or citizens' initiatives.

Administrative enforcement mechanisms apply primarily to political advertising and data protection. Under Article 31¹ of the Pre-election Agitation Law, the head of KNAB is empowered to issue binding administrative orders prohibiting the further dissemination of non-labeled or deceptive AI political campaign assets following expert evaluation. Decisions issued by KNAB are subject to judicial review before the administrative regional court, though filing an appeal does not suspend the execution of the removal order.

Data Protection Framework

Data processing within AI systems in Latvia is governed by the General Data Protection Regulation (GDPR) and national data protection law, overseen by the Data State Inspectorate (Datu valsts inspekcija, DVI). The Law on the Latvian Artificial Intelligence Centre introduces explicit statutory data protection rules for the AI Centre and its special regulatory environment. Personal data requested from state information systems for AI research or testing must adhere strictly to principles of data minimisation, mandatory anonymisation or pseudonymisation where feasible, and restricted access limited exclusively to directly involved entities under formal written agreement. All personal data used within the sandbox environment must be permanently deleted upon project completion.

Complementing these statutory rules, the draft National Data Governance Strategy (Nacionālā datu pārvaldības stratēģija) builds upon the Digital Transformation Implementation Plan. The strategy establishes the national data distribution and governance platform (DAGR), managed by the State Regional Development Agency (VRAA), to provide standardized metadata, API interoperability, automated journaling, and transparent audit trails for public-sector data sharing and personal data re-use.

Sector-Specific Rules

Sectoral rules targeting AI deployment are currently focused on the political, electoral, and public administration domains. The Pre-election Agitation Law lays down strict operational requirements prohibiting automated systems from using fake or anonymous social media profiles to conduct pre-election agitation, alongside transparency duties for paid synthetic content. In public administration, policy guidelines require state institutions to assess safety, privacy, and cost-efficiency when deploying virtual assistants, analytical tools, or automated processing systems.

Regarding healthcare AI, financial algorithms, employment tools, and autonomous vehicles, the corpus holds no record of standalone sector-specific AI statutes; instead, these domains remain governed by general legal frameworks and existing regulatory authorities. However, the AI Centre is statutorily tasked with encouraging safe AI adoption across key national sectors, including defence, internal security, health, education, culture, and sports.

International Alignment

Latvian AI regulation and strategic planning are closely aligned with European Union legislative initiatives and international benchmarks. The statutory mandate of the Artificial Intelligence Centre explicitly positions the institution to facilitate alignment with the EU AI Act, support national conformity assessment infrastructure, and participate in cross-border standardisation and research efforts. National policy planning documents incorporate indicators derived from European Commission frameworks and OECD AI Principles.

Latvia's data governance initiatives, including the National Data Governance Strategy, directly integrate requirements from European instruments such as the EU Data Governance Act, NIS2 Directive, and GDPR. By participating in EU digital programs and aligning sandbox mechanisms with European guidelines, Latvia ensures interoperability with broader European AI governance structures.

Future Developments

Latvia's regulatory timeline includes several upcoming implementation milestones. Under the Law on the Latvian Artificial Intelligence Centre, the Cabinet of Ministers is required to issue detailed secondary regulations governing the special regulatory environment and specific personal data processing protocols by May 31, 2025. Furthermore, beginning June 1, 2026, the AI Centre is set to assume its statutory role of issuing formal opinions on AI usage in election campaigns and electoral processes.

Progress continues on the draft National Data Governance Strategy, with planned delivery targets through 2026–2027 for technical components of the national DAGR platform. These developments aim to expand controlled access to high-value public datasets while ensuring automated logging, privacy controls, and auditability for advanced AI applications.

Enforcement Bodies

Agency	Mandate	Key Powers	Website
Corruption Prevention and Combating Bureau (KNAB)	Enforces pre-election agitation regulations, campaign financing, and political advertising restrictions.	Issues binding administrative orders to prohibit and remove non-compliant or non-labeled AI campaign materials after expert evaluation.
State Police (Valsts policija) and Prosecutor's Office (Ģenerālprokuratūra)	Investigates and prosecutes criminal offences under the Criminal Law.	Conducts criminal investigations, evidence preservation, digital forensics, and criminal prosecution for deepfake election manipulation under Article 90.1.
Data State Inspectorate (Datu valsts inspekcija - DVI)	Supervises personal data protection compliance and privacy rules across public and private sectors.	Enforces GDPR and national personal data processing law, investigates data protection violations, and monitors platform logging and privacy safeguards.
Ministry of Smart Administration and Regional Development (VARAM)	Coordinating ministry for digital transformation, national data policy, and statutory supervisor of the AI Centre foundation.	Formulates digital policy, submits Cabinet regulations, coordinates strategic planning, and exercises administrative oversight over delegated state tasks.