Latvia - AI Centre Establishment

Overview

The Law establishes the foundation "Artificial Intelligence Centre" (the Centre) as a national instrument to create and operate an AI technology ecosystem in Latvia. Its core legal objectives are to foster cooperation among public authorities, private sector actors and higher education institutions; to promote responsible AI adoption and innovation; and to mitigate risks to human rights, democratic processes and public security. The Law provides the Centre with corporate status (a foundation registered under the Associations and Foundations Law), authorises State budget funding and foreign donations, and tasks the Centre with project selection, funding, dataset curation and guidance production. It also creates a statutory route for a special regulatory environment — a controlled testing and development space with temporary derogations from certain legal requirements where necessary — subject to Cabinet regulations. The full enacted text and official publication are available on the national legal portal and the official gazette. (Law text, Latvijas Vēstnesis)

Definitions

The Law uses statutory definitions and references to established legal concepts. The "Centre" is defined as a foundation (nodibinājums) registered in the Register of Associations and Foundations; "special regulatory environment" denotes a temporary administrative regime for AI development and testing where certain otherwise‑applicable regulatory requirements may be relaxed to an objectively necessary extent; "personal data" follows the terminology used in Latvian and EU data protection law and is to be processed in line with minimisation, anonymisation/pseudonymisation principles and limited access. Transitional provisions set clear deadlines for registration, commencement of certain functions and the issuance of Cabinet regulations. See the consolidated text for full definitional context (official law text).

Governance and Institutional Framework

Governance arrangements are multi‑level: the Centre’s statutory structure comprises a Supervisory Board (nine members), a Director (as the Management Board), and a Secretariat function provided by the State Digital Development Agency. The Supervisory Board performs supervisory and advisory functions and appoints/dismisses the Director; its composition is explicitly mixed — representatives from the Ministry of Smart Administration and Regional Development, the Ministry of Economics, the Ministry of Defence, private sector and higher education are included — to ensure cross‑sectoral oversight and strategic alignment. The Secretariat, executed by the State Digital Development Agency, is tasked with operational support (IT and telecoms, legal, procurement, records, information security); this positions an existing state agency to supply continuity and ensure compliance with public information security standards. For delegated State administration tasks, the Centre is under the supervision of the Ministry of Smart Administration and Regional Development. The statutory design balances public oversight with private and academic participation to promote partnership while retaining accountability. (Law text, Saeima press release)

Key Focus Areas

The Centre’s remit covers multiple mutually reinforcing focus areas: (1) research, innovation and funding facilitation — identifying and coordinating national and international projects, running tenders and raising funds; (2) risk identification and resilience — proactively identifying security and safety risks in AI systems, including those that may undermine democratic processes (e.g., deepfakes) and national security, and promoting mitigation capabilities; (3) ethical and rights‑respecting deployment — ensuring AI is used ethically and respecting fundamental human rights; (4) skills and inclusion — increasing societal skills, accessibility and equality in AI; (5) data stewardship — curating and organising datasets for training AI with explicit safeguards; (6) national cultural and language resilience — promoting Latvian language and cultural data inclusion in AI solutions; and (7) regulatory experimentation — administering a special regulatory environment for safe testing and validation of systems with time‑bound derogations where objectively necessary. These focus areas link to sectoral objectives including health, education, defence, public administration, and national competitiveness. The Law also tasks the Centre with providing opinions on AI usage in election campaigns and elections, reflecting a preventative approach to information integrity. (Law text, LSM article)

Implementation Framework

Operationalisation is driven by a combination of Centre governance acts (statutes, supervisory decisions, administrative acts for the special regulatory environment) and Cabinet‑level secondary legislation that the Law requires by set deadlines. The Cabinet must publish implementing regulations governing the special regulatory environment and detailed data processing and technical requirements. The Ministry of Smart Administration and Regional Development is responsible for establishing the Centre and registering it in the Associations and Foundations Register by 31 March 2025; the Centre may then run tenders, receive funding, form partnerships and operate the special environment in accordance with Cabinet rules. The Secretariat provided by the State Digital Development Agency supplies key day‑to‑day functions, ensuring that procurement, IT and information security meet public standards. The Law requires openness and transparency in acquisition and use of resources and prescribes contractual safeguards (e.g., termination of free‑use transfers if contractual terms are breached). (Law text, official gazette)

Monitoring and Evaluation

Monitoring is multi‑faceted: internal oversight comes from the Supervisory Board and the Director, operational oversight for delegated state tasks comes from the supervising ministry, and adherence to administrative law provides judicial review of administrative acts issued by the Centre (appeals under the Administrative Procedure Law). The Law obliges transparency in financial and property usage and prescribes that administrative acts in the special regulatory environment take into account recommendations and objections of competent institutions. For the special environment, selection and evaluation procedures for projects must be prescribed by the Cabinet, enabling formalised selection criteria, evaluation metrics and termination conditions that can be audited. Data processing within the environment is subject to documented agreements and technical/organisational safeguards, facilitating oversight and post‑project deletion or archival supervision. (Law text)

Penalties, Liability, and Appeals

The Law does not primarily establish new criminal or administrative fines but creates legal mechanisms with enforcement consequences: exclusion from participation in the special regulatory environment (for sanctioned entities, those under specified foreign influence, or merchants with recent criminal sanctions), contractual termination clauses (e.g., termination of free‑use property transfers for non‑compliance), and court remedies — administrative acts issued by the Centre may be appealed to courts in accordance with the Administrative Procedure Law. The Law also references compliance with other Latvian statutes that carry separate enforcement powers (e.g., criminal or administrative sanctions under other laws for misconduct). The Centre’s obligations to ensure data deletion, minimal processing and restricted access create potential contractual and reputational liabilities for participants and the Centre. Judicial review and transparency obligations function as primary checks. (Law text)

Relationship to Other Instruments

The Law sits alongside and references multiple existing instruments: the Associations and Foundations Law (governing registration, governance and legal status of the Centre), the Law on Prevention of Squandering of Financial Resources and Property of Public Entities (governing transfers from state capital companies to the Centre), the Administrative Procedure Law (appeals and judicial review of administrative acts), and national data protection provisions and EU law (which shape personal data handling even where the Centre may request access to state information systems for research/public interest purposes). Cabinet regulations called for by the Law will operationalise how temporary derogations interact with existing sectoral regulatory frameworks. The Centre’s statutory right to prepare recommendations and participate in law and regulation development formalises its role as a policy‑shaping actor. (Law text, Associations & Foundations Law)

International Alignment

The Act explicitly authorises the Centre to participate in international organisations and partnerships and to attract foreign funding and donations. This enables alignment with EU‑level AI governance initiatives (including the EU AI Act framework), participation in research consortia, and cooperation with international partners on standards, interoperability and best practice. The Law’s emphasis on rights‑respecting use, transparency and risk mitigation supports alignment with international human rights and data protection obligations. The requirement that the Centre coordinate with competent institutions in creating the special regulatory environment also facilitates cross‑border regulatory dialogue and alignment on testing conditions and safeguards. (Law text)

Implementation Timeline

Sources and References