Reglamento de la Ley 31814, high-risk AI system duties
A citation is an address, not a summary. The first part names the law; what follows narrows it to the exact section, article or paragraph.
In force 8 months, effective 22 January 2026.
An AI risk obligations rule binding public and private bodies.
As of 5 September 2026.
What it requires
- Before a high-risk AI system interacts with a user, tell them, clearly and simply, its purpose or use, its main functions, and the kind of decisions it can make.
- Where a high-risk AI system's decision affects a person's human rights, explain the result to the affected user in accessible language, covering the criteria and factors the decision relied on.
- Ask the Secretaría de Gobierno y Transformación Digital for a determination if you are unsure whether your system's use falls within the high-risk category.
- A private-sector developer or implementer's transparency duty phases in by sector over one to four years from the Decreto Supremo's publication, with a longer schedule for small businesses and innovative start-ups; confirm your sector's deadline before treating the duty as not yet binding.
- A breach of the personal-data and privacy rules that apply to a high-risk system is sanctioned under Ley 29733's own regime, not under this Decreto Supremo.
If you get it wrong
Criminal exposureNo
Private right of actionNo
Who enforces it
Enforcement body
Secretaría de Gobierno y Transformación Digital (SGTD); a breach of the personal-data rules a high-risk system must follow is instead enforced by the Autoridad Nacional de Protección de Datos Personales under Ley 29733
What it reaches
Obligation class
Disclosure, Governance
What this law does
Drafted with AI from the cited sources under the direction of UnGovr staff. UnGovr holds editorial responsibility for this page.
Article 24 classifies an AI-based system's use as high-risk where, among other criteria, it manages critical national assets supporting essential services (energy, telecommunications, health, transport, water, and banking, among others) or is used to evaluate a person in a selection process. A developer or implementer may ask the SGTD to determine whether a use falls within the high-risk category.
Article 25 requires the developer or implementer of a high-risk system to establish mechanisms guaranteeing algorithmic transparency, informing the user beforehand, clearly and simply, of the system's purpose or use, its main functions, and the kind of decisions it can make, and, where the system's decisions affect human rights, to explain its results to affected users in accessible language.
For a private-sector developer or implementer, article 25's transparency duty and a related Title VI chapter phase in gradually by sector, starting the day after the Decreto Supremo's publication: one year for AI uses in health, education, justice, security, and economy-finance; two years for transport, commerce, and labor; three years for production, agriculture, energy, and mining; and four years for every other use nationwide, with a further, longer schedule for small businesses and innovative start-ups.
Public-sector entities implement article 25 on their own graduated schedule, running from one to three years after the Decreto Supremo's publication depending on the type of entity; only for the smallest local governments (Tipo D, E, F, and G) is implementing article 25 described as facultative, according to their resources and capacities.
Article 26 requires compliance with the personal-data and privacy rules in force for any high-risk system's development, implementation, or use, and states that liability for a breach of those data-protection duties is determined under Ley No. 29733's own sanctioning regime rather than under this Decreto Supremo.
When LexLint raises it
high_risk_decisionsautomated_outreach