Chile’s AI Policy Is In Force: What Teams Should Do

Chile’s National Policy on Artificial Intelligence is in force, giving public bodies and businesses a national framework for responsible AI. That status does not mean a new AI law has just taken effect: the policy became effective on December 3, 2021, and was formally updated on January 28, 2025.

What's changing

The practical message is that Chile’s AI policy is an active framework, not a proposal awaiting implementation. Its 10-year roadmap guides the ethical, inclusive and sustainable development and adoption of AI across the public and private sectors.

The timeline matters. Decreto Supremo N°20 was issued on September 20, 2021, and the policy was publicly presented on October 28, 2021. Publication in the Diario Oficial followed on December 3, 2021, the effective date. Consultations and update work took place during 2023–2024, including work on governance and ethics. The formal update was published through Decreto N°12 on January 28, 2025. None of those dates is an approaching deadline.

The National Policy on Artificial Intelligence takes a broad view of AI: computational systems performing tasks such as prediction, recommendation or decision-making, with different degrees of autonomy. Teams should therefore look beyond chatbots when considering its relevance. Predictive tools and automated recommendation systems also fit within that broad description.

Its core expectations are human-centered design, respect for rights, inclusion, non-discrimination, transparency and accountability. It also promotes risk mapping and impact assessments, particularly for high-impact or high-risk systems. Documentation practices such as model cards and data statements can help explain how systems work and what limitations users should understand.

Crucially, this is a policy, not a standalone law imposing new penalties. It does not itself introduce fines or criminal sanctions. The policy is active now, but its recommendations should not be presented as universally binding compliance duties. Existing sectoral laws may still apply independently, and future AI rules may draw on the policy’s risk-based direction.

Who is affected

This is a national policy for Chile, not a measure limited to a particular region or municipality. Its scope covers public and private involvement in AI research, development, deployment and public use.

That makes it relevant to researchers, founders, product managers, developers, public officials and in-house teams selecting or operating AI tools. Its practical reach is not confined to large technology companies. Smaller businesses using third-party systems should also consider whether their choices align with its principles. The supplied policy information does not establish a company-size threshold or a small-business exemption.

Public-sector suppliers have a particular reason to pay attention. The policy advocates integrating AI criteria into public procurement, potentially influencing how government buyers assess systems and vendors. That is a policy direction, not evidence that every tender already contains identical AI requirements. Suppliers should check the actual terms of each procurement process.

The Ministry of Science, Technology, Knowledge and Innovation coordinates implementation, involving an interministerial committee and multi-actor working groups. For organizations, this means implementation can develop through several channels rather than through a single enforcement process created by the policy.

Three things to do this week

1. Inventory your AI and identify its highest-impact uses. List systems your organization develops, buys or deploys, including prediction and recommendation tools. Record each system’s purpose, owner, users and people affected. Then prioritize a risk review for uses where errors, bias or limited human oversight could have serious consequences. This puts the policy’s risk-mapping and impact-assessment recommendations into practice without assuming every system carries the same risk.

2. Build a usable documentation baseline. For one priority system, prepare a model card or similar system record and a data statement where appropriate. Describe intended use, limitations, data sources, evaluation results and available human oversight. Assign someone to maintain that documentation. These are practical ways to support the policy’s transparency and accountability goals; they are not a claim that the policy mandates a particular template.

3. Check procurement terms and existing legal duties. If you sell to government, review current tender requirements and prepare evidence explaining your AI system’s risks and safeguards. Separately, identify applicable data-protection, consumer-protection or sector-specific duties. The absence of new penalties in the policy does not remove obligations under existing law. Keep that legal review distinct from voluntary alignment with the policy.

Related context

Read the policy alongside the Action Plan of the National AI Policy for implementation context. The formal policy update approved through Decreto N°12, published in January 2025, is also relevant to understanding its development after consultation.

For public-sector applications, the MinCiencia instructive on AI use in government provides related context. These instruments should be read on their own terms: a policy, an action plan and government instructions are not interchangeable sources of binding duties.

The immediate takeaway is to align AI governance with Chile’s active national direction while checking separately which legal and contractual requirements actually bind your organization.

Note: this article was drafted by AI - OpenAI GPT-6.1