Latvia - AI Use in Campaigning (2024)

Amendments to the Pre‑election Agitation Law regulating use of AI in campaigning

Grozījumi Priekšvēlēšanu aģitācijas likumā

Latvia

RAI-LV-NA-APARUXX-2024
Effective: November 7, 2024
In Force(In Force)
ActTransparency and DisclosureGovernance and OversightEnforcement and Penalties
Export PDF

In October–November 2024 the Saeima adopted amendments to the Pre‑election Agitation Law that require paid pre‑election materials created with artificial intelligence (AI) to be clearly and unambiguously labelled, ban the use of fake or anonymous social‑media accounts driven by automated systems for political agitation, and grant the Corruption Prevention and Combating Bureau (KNAB) authority to order removal of AI‑generated campaign content. The law entered into force on 7 November 2024.

Summary

Background and purpose: The 2024 amendments to Latvia's Priekšvēlēšanu aģitācijas likums (Pre‑election Agitation Law) were prompted by concerns that AI‑generated audio, image and video content (including deepfakes and fabricated events) could be used to mislead voters during election periods. The initiative originated from the Office of the President and was debated in Saeima committees; it was adopted in the final reading on 24 October 2024 and promulgated by the President on 6 November 2024, entering into force on 7 November 2024. Core obligations: The amendments add a new Article 3¹ which requires that, during the pre‑election agitation period, any paid pre‑election agitation or paid campaign materials that include an AI‑system generated depiction of a person or an event that is not true to reality (images, audio or video) must be clearly and unambiguously identified as such. The labelling duty is aimed specifically at paid campaign materials rather than all organic social activity. Prohibitions and enforcement: The revised Article 18 introduces a prohibition on using automated systems to perform pre‑election agitation through fake or anonymous social‑media account profiles. The amendments also add Article 31¹ granting the head of the Corruption Prevention and Combating Bureau (Korupcijas novēršanas un apkarošanas birojs, KNAB) the power to adopt a decision prohibiting further dissemination (placement) of AI‑generated pre‑election agitation materials where an infringement is found. KNAB must obtain an expert or other specialist opinion and assess proportionality and feasibility before issuing such removal orders. KNAB is required to immediately publish any decisions to adopt, revoke or amend such prohibitions on its website. Judicial review: Decisions of the KNAB head under Article 31¹ may be appealed to an administrative regional court; filing an appeal does not suspend the KNAB decision. Scope and interplay with other law: These amendments operate alongside other election and media rules in Latvia and interact with EU‑level obligations (notably the EU AI Act and GDPR). The law focuses on transparency of AI usage in paid campaigning, curbs automated anonymous influence campaigns, and establishes an administrative removal mechanism rather than a broad criminal prohibition. Practical implications for stakeholders: Political parties, candidates, campaign consultants, political advertisers, media sellers and online platforms that accept paid pre‑election placement must ensure labelling of AI‑generated imagery, audio and video when used in paid materials; they must also avoid using automated fake or anonymous profiles for agitation. KNAB will monitor compliance and may issue take‑down orders. Relationship to EU law: The amendments align with European efforts to regulate AI in sensitive contexts, and with the EU AI Act's emphasis on transparency obligations for certain AI outputs and the prohibition of specific manipulative practices. Implementation and compliance: Practically, entities will need clear internal policies and audit trails showing when AI was used, and labelling for end‑users; platforms and advertisers should update their advertising contracts and placement policies; KNAB will develop procedures to request expert opinions and to publish decisions. The measure aims to protect the integrity of electoral discourse and voters’ right to receive truthful information while leaving room for contestation via administrative and judicial remedies.

Full article

Read full text ↗

Overview

The 2024 Amendments to the Pre‑election Agitation Law ("Grozījumi Priekšvēlēšanu aģitācijas likumā") were adopted by the Saeima on 24 October 2024 and promulgated by the President on 6 November 2024; they entered into force on 7 November 2024. The package inserts a new Article 31 imposing mandatory labelling where AI systems produce paid pre‑election content depicting a person or fabricated event (image, audio or video), prohibits automated fake or anonymous social‑media accounts under Article 18(6), and grants the Korupcijas novēršanas un apkarošanas birojs (KNAB) authority to order the removal of AI‑generated agitation material under Article 311. The official consolidated text is published and maintained on the national legislation portal Likumi.lv: Grozījumi Priekšvēlēšanu aģitācijas likumā, and the Saeima summary is available at Saeima: Voters must be informed if AI was used.

Definitions

The amendments rely on operational definitions embedded in the law and practice: "artificial intelligence system" is used functionally to cover systems that generate images, audio or video, including synthetic person depictions and fabricated events; "paid pre‑election agitation" denotes advertising or other remunerated campaign materials during the official pre‑election period; "person depiction" covers generated imagery or synthetic audio purporting to represent a real individual; and "real‑world inconsistent event" covers media that show events that did not occur but are presented as if they did. For enforcement, KNAB may obtain expert opinion to assess whether content meets these legal definitions.

Governance and Institutional Framework

Primary enforcement authority is the Korupcijas novēršanas un apkarošanas birojs (KNAB), which is empowered to evaluate alleged breaches and issue binding administrative decisions to prohibit further distribution of non‑labelled AI‑generated paid agitation material. KNAB must solicit an expert or other specialist opinion and consider proportionality and feasibility before issuing removal orders; it must then promptly publish the decision, its revocation or amendment on its website. The Saeima remains the legislative owner for any later changes; the Central Election Commission (Centrālā vēlēšanu komisija, CVK) retains its competencies regarding election administration. National publication and registration functions (official gazette) are fulfilled by Latvijas Vēstnesis; the consolidated text is hosted on Likumi.lv and the formal promulgation reference is recorded at Latvijas Vēstnesis, OP 2024/217.9. Data protection or privacy incidents remain subject to the Data State Inspectorate procedures and to GDPR where personal data processing is involved (Datu valsts inspekcija).

Key Focus Areas

The law focuses on three principal areas: (1) transparency and labelling — paid campaign materials containing AI‑generated person depictions or fabricated events must be "clearly and unambiguously" marked as AI‑generated; (2) automated anonymity and fake accounts — the law prohibits using automated systems to perform pre‑election agitation with fake or anonymous social media profiles; and (3) administrative removal and publication — KNAB may adopt removal orders after expert assessment and must publicise decisions. These elements reflect European policy priorities (integrity of information during elections, transparency of synthetic content and measures against automated influence operations) and are intended to curb deceptive manipulation without banning lawful political speech. The labelling obligation applies to paid placements, capturing political advertising and paid promotional content; organic unpaid posts are not the primary target, though platform policies and other laws (e.g., electoral financing rules) may interact.

Implementation Framework

Obligations fall on political organizations, campaign advertisers, media and advertising intermediaries, and platforms selling paid space: they must identify paid materials that incorporate AI‑generated person depictions or fabricated events and include an explicit label that the material was AI‑generated. In practice this requires internal procedures to document use of AI tools (provenance records, timestamps, model identifier where available), explicit labelling on media assets (overlays, preambles or disclaimers), and contract clauses with service providers and media sellers to ensure labelling upstream. Platforms that host paid political advertising should update their advertising policies to reflect the statutory labeling requirement and to prevent placement of paid ads that use fake/anonymous automated accounts. KNAB will define its operational approach to seek expert opinions and set standards for proportionality before issuing removal decisions; guidance and administrative instructions are expected to be prepared following the law's entry into force.

Monitoring and Evaluation

Monitoring will be primarily administrative: KNAB will receive complaints and signals, may proactively monitor paid political ads, and must publish its decisions to ensure transparency and public oversight. The amendments institutionalize the publication duty so that stakeholders and researchers can track enforcement patterns. Evaluation metrics will likely include the number of notices issued, take‑down orders, appeals lodged, and the time taken to secure expert opinions. Interaction with platform reporting channels and media transparency registers (where present) will be important for effective monitoring. Periodic reviews by parliamentary committees or independent oversight bodies are foreseeable to assess whether the narrow labelling and removal regime adequately protects electoral integrity.

Penalties, Liability, and Appeals

The amendments themselves create an administrative removal power for KNAB; they preserve the right to judicial appeal in the Administrative Regional Court against KNAB decisions, and appeal does not suspend the KNAB order. The statutory text contemplates application of other administrative or criminal liability where relevant (for example, in cases of fabricated material that also meets criminal thresholds such as impersonation or fraud), and existing administrative penalties under the Pre‑election Agitation Law remain available for breaches of campaign rules. Key procedural safeguards include the required expert opinion and the proportionality assessment before issuing removal orders. Entities subject to KNAB decisions can pursue remedies through administrative court review and, ultimately, higher courts if necessary.

Relationship to Other Instruments

The amendments operate alongside multiple legal frameworks: the core Pre‑election Agitation Law structure, electoral financing and advertising rules, the Data Protection Regulation (GDPR) as applied domestically through the Data State Inspectorate, and criminal law provisions that may govern forgery, fraud or identity offences. At the EU level, the provisions complement obligations and prohibitions anticipated by the EU’s horizontal AI Act framework (transparency for certain AI outputs and prohibitions for manipulative systems) and the e‑privacy and platform governance initiatives. National guidance or implementing administrative rules will be needed to reconcile labelling mechanics, platform notice‑and‑takedown processes and cross‑border content distribution challenges.

International Alignment

Latvia’s amendments reflect an EU‑wide trend to require transparency about generated content in political contexts and to empower regulators to address harmful AI uses in elections. They are consistent with the direction of the EU AI Act and with member‑state measures that focus on deepfakes, synthetic media transparency, and measures against inauthentic coordinated behaviour. The law is narrowly tailored to paid pre‑election messaging and to administrative removal, reducing the risk of conflict with free speech protections and providing a model for other jurisdictions seeking targeted transparency and administrative enforcement powers. Cross‑border cooperation (platform contacts, information exchange, mutual legal assistance) will be important where content originates outside Latvia or is hosted by foreign platforms.

Implementation Timeline

EventDate
Saeima adoption (final reading)2024‑10‑24
Presidential promulgation2024‑11‑06
Entry into force (next day)2024‑11‑07
Expected KNAB operational guidance (administrative)Within 3–6 months after entry into force
First cycle review (parliamentary / administrative review)12 months after entry into force

Sources and References

SourceType
Grozījumi Priekšvēlēšanu aģitācijas likumā (Likumi.lv)Primary Source
Saeima press note: Voters must be informed if AI was usedPrimary Source
Latvijas Vēstnesis publication OP 2024/217.9Primary Source

Requirements for a company

What an organisation has to do under Latvia - AI Use in Campaigning (2024), at a glance. Not legal advice — the table below gives the provision and deadline for each item.

Must do

6
  • Clearly label all paid pre-election materials created with AI depicting a person or fabricated event.Political organizations, campaign advertisers, media, and advertising intermediaries.
  • Do not use automated systems for pre-election agitation with fake or anonymous social media profiles.Anyone performing pre-election agitation.
  • Prevent the placement of paid political ads that use fake or anonymous automated accounts.Platforms hosting paid political advertising.
  • Document the use of AI tools for creating paid pre-election materials.Political organizations, campaign advertisers, media, and advertising intermediaries.
  • Include clauses in contracts with service providers and media sellers to ensure AI labelling.Political organizations and campaign advertisers.
  • Update advertising policies to reflect the statutory AI labelling requirement.Platforms hosting paid political advertising.

Must not do

0

Nothing in this category.

Should do

0

Nothing in this category.

Should not do

0

Nothing in this category.

Who must do what

The obligations under Latvia - AI Use in Campaigning (2024), most serious first. Not legal advice — verify against the official text before relying on it.

#WhoRequirementBy whenWhereSeverity
1Political organizations, campaign advertisers, media, and advertising intermediaries.Clearly label all paid pre-election materials created with AI depicting a person or fabricated event.
Article 3^1 imposing mandatory labelling where AI systems produce paid pre‑election content depicting a person or fabricated event
Before placing on marketArticle 3^1Critical
2Anyone performing pre-election agitation.Do not use automated systems for pre-election agitation with fake or anonymous social media profiles.
prohibits automated fake or anonymous social‑media accounts under Article 18(6)
Nov 7, 2024Article 18(6)Critical
3Platforms hosting paid political advertising.Prevent the placement of paid political ads that use fake or anonymous automated accounts.
prevent placement of paid ads that use fake/anonymous automated accounts.
Nov 7, 2024Article 18(6)Critical
4Political organizations, campaign advertisers, media, and advertising intermediaries.Document the use of AI tools for creating paid pre-election materials.
requires internal procedures to document use of AI tools (provenance records, timestamps, model identifier where available)
Before placing on marketImportant
5Political organizations and campaign advertisers.Include clauses in contracts with service providers and media sellers to ensure AI labelling.
contract clauses with service providers and media sellers to ensure labelling upstream.
Before engaging service providers or media sellersImportant
6Platforms hosting paid political advertising.Update advertising policies to reflect the statutory AI labelling requirement.
Platforms that host paid political advertising should update their advertising policies to reflect the statutory labeling requirement
Important

© Regulations.AI · updated on 13-Jun-2026