Jump to content

Dominican Republic Public Services AI Ethics Code (Decree 787-26)

From AI Law Wiki

AI-generated text. This page was generated using artificial intelligence.

The Dominican Republic's Decree 787-26, signed on November 2, 2026, approves the Código de Ética de la Inteligencia Artificial para los Servicios Públicos and incorporates the code as an annex. It establishes responsibilities for public bodies that develop, acquire, operate or use AI in public services and administrative functions.[1]

Scope and institutional responsibilities

The decree covers centralized and decentralized public-administration entities, including public companies. Its functional scope includes systems supporting administrative acts and routine staff assistance used in covered public functions. Contracting institutions must include compliance, data protection, security, auditing, traceability and supervision requirements in procurement documents and contracts.[2]

The Ministry of Public Administration has administrative-policy responsibilities. The Directorate General of Government Ethics and Integrity (DIGEIG) oversees ethics and integrity, including review of ethics and rights impact assessments. The Government Office of Information and Communication Technologies (OGTIC) supplies technical standards and support. The decree also establishes a national AI ethics council for consultation and coordination. Agency authorities remain responsible for implementation and must identify accountable units and people for their systems.[3]

Assessment, transparency and procurement

The code requires an ethics and rights impact assessment (EIED) before production deployment of systems affecting rights, essential services, public resources or sanctions. High-impact systems require a completed assessment, review by the designated national body and publication of an accessible summary before deployment. Assessment covers purpose, necessity, alternatives, risks, mitigation and human oversight, including environmental and community effects. Continuing monitoring can lead to correction, restrictions, suspension or withdrawal.[4]

Institutions must maintain system inventories and publish periodic extracts subject to justified privacy and security limits. People affected by significant AI-supported administrative decisions must receive understandable information about its use and principal criteria and have channels for explanation, human review and complaints. Procurement safeguards include government control of data and rights to audit and obtain explanations. Commercial secrecy must not prevent the verification needed for ethical, technical and legal compliance.[5]

Restricted uses and children

The code prohibits general social scoring that determines access to rights or public services and autonomous penal or coercive decisions without human involvement and due process. It restricts indiscriminate mass biometric identification in public spaces, requiring a specific judicial order, legitimate purpose, adequate legal basis and independent oversight for an exception.[6]

AI whose principal target audience or foreseeable users are children carries a reinforced presumption of risk. Public-service use requires an approved, specific impact assessment showing a legitimate purpose in children's best interests, necessity, no less intrusive alternative, exceptionally low residual risk, specific safeguards, continuing human supervision and accessible remedies. Protective tools are subject to those conditions. The code separately prohibits AI used to generate, facilitate or distribute harmful content involving minors, including synthetic sexualized or abusive material.[7]

Transition and enforcement

Article 14 gives institutions 180 calendar days from publication to make institutional adjustments, inventory systems and designate responsible units. Existing high-impact systems must undergo assessment within the period set by DIGEIG, with human supervision maintained during adjustment. The exact gazette publication date and resulting calendar deadline have not been verified in this coverage.[8]

The code relies on existing public-service integrity and administrative-law mechanisms rather than creating an autonomous sanctioning regime. It requires responsible officials, training, monitoring and complaint handling, with participatory review at least every two years.[9]

References

  1. ↑ Decree 787-26 and annexed public-services AI ethics code, signed November 2, 2026, official Ministry of Culture copy, decree articles 1–3 and 16. The annex is marked version 1.0, July 2026.
  2. ↑ Decree 787-26, articles 2–3, in the official decree and annex.
  3. ↑ Decree 787-26, articles 4–11 and 13, in the official decree and annex.
  4. ↑ Annexed code, section 7, printed pages 30–31 (PDF pages 44–45), in the official decree and annex.
  5. ↑ Decree 787-26, article 11; annexed code, sections 8–10 and 12, in the official decree and annex.
  6. ↑ Annexed code, section 11, printed pages 33–34 (PDF pages 47–48), in the official decree and annex.
  7. ↑ Annexed code, section 11.1, printed page 35 (PDF page 49), in the official decree and annex.
  8. ↑ Decree 787-26, article 14, in the official decree and annex; Executive Legal Consultancy announcement, November 4, 2026.
  9. ↑ Annexed code, sections 13–16; decree article 10, in the official decree and annex.
Retrieved from "https://ailawwiki.com/index.php?title=Dominican_Republic_Public_Services_AI_Ethics_Code_(Decree_787-26)&oldid=8328"