/ Bill Regulating Artificial Intelligence Introduced to Congress

Companies operating within the scope established by the bill will be required to thoroughly assess their AI systems and implement only those systems that are not prohibited and comply with the requirements and standards set out in each risk classification.

Bárbara Silva Ordóñez
Associate Attorney

On May 7, 2024, the President of the Republic presented to the Chamber of Deputies a bill to regulate artificial intelligence (AI) in Chile (Bill No. 16821-19), aimed at regulating the use of AI systems to promote development, utilization, and adoption of this technology while safeguarding the fundamental rights of individuals.

The recipients of this bill are the providers who introduce AI systems into the market or make them available for use in the national territory; implementers, importers, and distributors of AI systems, when they or their authorized representatives are domiciled in the national territory; and providers and implementers of AI systems domiciled abroad when the output information of the AI system is used in Chile (Art. 2).

Following the line of the AI Act (Regulation approved in March 2024 by the European Commission that creates the regulatory framework for AI systems in the European Union), the bill adopts an approach based on the level of risks affecting fundamental rights.

It encompasses four categories of AI systems, analogous to those of the European regulation:

1.- Unacceptable Risk

Incompatible with the respect and guarantee of the fundamental rights of individuals, so their introduction into the market or deployment is prohibited. Examples include the creation of deepfakes with sexual content exploiting vulnerabilities of children and real-time biometric identification for purposes other than public safety.

2.- High Risk

Capable of negatively affecting the health and safety of individuals, their fundamental rights, or the environment, as well as consumer rights, especially if they fail or are used improperly. An example is the use of AI in recruitment and hiring processes, where it could evaluate and filter job applications, potentially influencing the employment opportunities of applicants.

3.- Limited Risk

Their use presents a non-significant risk of handling, deception, or error due to their interaction with individuals. For instance, a service company’s chatbot that responds to specific user queries within its area of expertise without impacting user safety or rights.

4.- No Obvious Risk

All those that do not qualify in the previous categories. An example is an application that suggests TV series based on user preferences.

The oversight and regulatory compliance are entrusted to the Personal Data Protection Agency, an entity that will be created following the approval of the bill amending the Privacy Protection Law (which is in its final legislative stage and provides for a legal vacancy of two years). The bill considers severe, serious, and minor infractions. Fines can reach up to 20,000 UTM (Arts. 24-27).

The bill represents a significant step forward in AI regulation in Chile, aligning with international standards such as those set out in the AI Act. This initiative not only promotes the development and adoption of AI technologies but also ensures the protection of fundamental rights through a well-defined risk classification.

Bill available at the following link: https://www.camara.cl/verDoc.aspx?prmID=17048&prmTIPO=INICIATIVA