We invite you to read the opinion piece by our Director of New Technologies and AI, Juan Pablo González, on the challenges posed by the regulation of artificial intelligence in Chile.
Chile is currently debating one of the most important issues of our time: how to use artificial intelligence systems responsibly. In this regard, it is crucial to address a series of gaps that currently exist in the regulatory framework.
The first of these is a lack of clarity regarding rights and responsibilities throughout the AI system development chain—that is, from the developers to the final product. Furthermore, the obligations to explain, audit, and, consequently, challenge decisions made by AI systems are unclear, except for regulations regarding personal data; this undoubtedly represents a shortcoming from a risk-based perspective and, therefore, limits the adoption of a preventive approach in this regard.
In terms of intellectual property, issues such as the authorship of AI outputs; the use of protected works and potential exceptions for training models; clearly defined liabilities for infringements; and, consequently, ownership of key components in model training processes have not been fully resolved at the regulatory level.
Furthermore, from a confidentiality perspective, there is no uniform regulation regarding the use of information on external platforms, rules for retaining or deleting information used to train models, or obligations of third parties—factors that can undoubtedly impact the ability to make reliable decisions based on the results of AI systems.
In the search for solutions, we can look to best practices from other countries. Standards such as ISO 42001 on AI Management Systems enable progress toward clear governance and, consequently, the definition of internal practices regarding AI systems and their use.
In this regard, there are a number of practices that can accelerate the safe adoption of artificial intelligence within organizations. These include creating an inventory of systems within the organization; managing the risks associated with AI throughout its entire lifecycle; and adopting a “privacy and security by design and by default” approach.
Likewise, transparency and traceability mechanisms are essential, as they will allow for training models through potential changes in system configuration; implementing controls ranging from access restrictions to human oversight; and establishing clear audit frameworks for AI systems.
The challenge is that we are trying to establish rules for a technology whose evolution and scope we cannot yet fully grasp. Therefore, regulating artificial intelligence cannot consist solely of addressing the risks we already know about, but rather of building an institutional framework capable of identifying, assessing, and managing those that will emerge. Chile has the opportunity to close current gaps by combining regulation, international standards, and governance mechanisms that allow for foresight without hindering innovation.
Column written by:
Juan Pablo González | New Technologies and AI Director | jgonzalez@az.cl
