Chile - Consumer Protection in AI (Resolución Exenta N°33/2022)

Overview

The SERNAC Interpretative Circular on AI (Resolución Exenta N°33, 18 January 2022) provides institutional guidance to Chilean providers and consumers about the legal and practical risks that the use of artificial intelligence systems may pose within consumer relationships. The document does not create a new law but clarifies how existing consumer protection and privacy rules apply to AI-driven processes. The Circular highlights the need for clear information, safeguards for data and non-discrimination, and preservation of consumer choice. The original SERNAC publication and the full text of the circular are available from SERNAC (see SERNAC circular index page) and the downloadable PDF of the Circular itself (SERNAC Resolución Exenta N°33 PDF), which SERNAC uses as the primary reference for enforcement and guidance.

Definitions

The Circular adopts working definitions grounded in existing Chilean law and policy. It refers to "systems of artificial intelligence" as automated systems that, according to defined objectives, perform tasks such as predictions, recommendations or decisions that affect consumers. "Data" is read broadly to include numeric, alphabetic, photographic, acoustic or other formats linked to identifiable persons (consistent with Law N°19.628). The Circular treats data on behaviors, preferences and consumer habits collected or inferred by AI as particularly sensitive in commercial contexts, warranting heightened protection. It also distinguishes automated decision-making that merely supports a human actor from fully automated decisions that directly shape contractual offers, pricing, or eligibility for services.

Governance and Institutional Framework

The Circular situates SERNAC as the lead consumer protection authority responsible for interpreting the Consumer Protection Law (Ley N°19.496) as it applies to AI in markets. It indicates that SERNAC will use its powers for market surveillance, investigations, and corrective actions when AI practices affect consumers' rights and may coordinate with other public actors on data protection and AI policy. The document references the institutional context of national AI strategy and data protection reform: SERNAC published the Circular within the framework of Chile's broader AI policy objectives and in coordination with other State efforts to manage AI risks. SERNAC's public announcement and the Circular's posting (available at SERNAC news and the PDF here) indicate the Service's intent to use interpretative guidance as a supervisory tool while recognising that primary data-protection laws and consumer law remain the legal basis for enforcement. The Circular also notes potential collaboration with ministries and agencies involved in data protection and AI policy (for example, ministries mentioned in national AI policy summaries, such as the Ministry of Science; see MinCiencia summary).

Key Focus Areas

The Circular concentrates on concrete areas where AI intersects with consumer rights and where SERNAC expects providers to demonstrate reasonable safeguards. First, information and transparency: providers must disclose when AI is used, the purpose of the system, its role in decision-making or contract execution, and the categories of data processed. Second, data protection: compliance with the applicable data-protection framework is required, and AI-derived profiling of consumer behavior is treated as sensitive; providers should apply heightened security and minimization principles. Third, non-discrimination and fairness: systems must not produce arbitrary or unjustified differential treatment based on protected or irrelevant characteristics; providers should test for and mitigate biases. Fourth, freedom of choice and avoidance of manipulation: design choices should not coerce consumers into transactions or exploit vulnerabilities through dark patterns or manipulative personalization. Fifth, safety, robustness and technical measures: providers must adopt technical and organizational measures to ensure confidentiality, integrity and availability of data processed by AI and to reduce foreseeable harms from errors or anomalous system behaviors. Sixth, human oversight and redress: where AI affects contractual formation or performance, providers should maintain channels for human review and effective remedies for consumers when they are adversely affected.

Implementation Framework

SERNAC positions the Circular as an interpretative standard to be used in supervising compliance, guiding complaints handling, and informing corrective actions. Providers are expected to integrate AI governance into existing compliance programs: document AI use cases, implement impact assessments for high-risk applications in consumer relations, run bias and fairness testing, maintain logs and traceability for automated decisions, adopt privacy-by-design measures and offer clear consumer-facing disclosures. The Circular recommends (but does not mandate by itself) that providers conduct prior risk identification for AI systems that personalize offers, determine prices, score consumers, or automate contract offers and that they define procedures for escalation to human agents and for remedying wrong or harmful outcomes. SERNAC signals that failure to align AI practices with consumer law obligations may result in investigations and sanctions under the Consumer Protection Law.

Monitoring and Evaluation

The Circular expects ongoing monitoring of AI systems and iterative evaluation to detect biases, degradation of model performance, data leaks, or emergent harms. Providers should implement metrics and monitoring routines that capture fairness, accuracy, security incidents and consumer feedback. SERNAC may request documentation, logs or evidence of testing as part of information requests or investigations; the Circular thus functions as a baseline for record-keeping and audit readiness. In practice, monitoring encompasses pre-deployment testing, post-deployment performance tracking, and procedures for consumer complaints and remediation.

Penalties, Liability, and Appeals

The Circular itself does not create new penalties but explains that non-compliance with consumer protection obligations — as interpreted in the context of AI — can lead to SERNAC enforcement actions under Ley N°19.496 and related rules. Potential administrative measures available to SERNAC include orders to comply, corrective measures, public warnings, and fines when statutory violation is established under applicable law. The document emphasises providers' civil liability to consumers for damages arising from defective performance or misconduct connected to AI-driven decisions and notes the importance of preserving evidence for liability and appeals. Consumers retain rights to administrative and judicial remedies under Chilean law.

Relationship to Other Instruments

The Circular cross-references domestic instruments that shape the legal environment for AI in consumer markets: the Consumer Protection Law (Ley N°19.496) for general consumer rights; Law N°19.628 (protection of private life / data protection) and its subsequent reforms for data handling; and the national AI policy context described by the Ministry of Science. SERNAC frames the Circular as complementary to upcoming or evolving data-protection reforms and national AI initiatives, recommending consistency between consumer protections and data governance frameworks to avoid regulatory gaps.

International Alignment

Although tailored to Chilean law, the Circular reflects internationally recognised principles for trustworthy AI: transparency, fairness, accountability, data minimization, security and human oversight. SERNAC referenced international practices and aligns its expectations with broader OECD and EU-style principles for AI governance. The Circular therefore serves both to translate international norms into a Chilean consumer-protection context and to encourage providers to apply international best practices when designing AI systems that interact with consumers.

Implementation Timeline

Sources and References