This month the first deadline for implementation of several of the obligations provided for in the Regulations of the Artificial Intelligence Law expired. The Executive, Legislative and Judicial Branch, constitutionally autonomous organizations and certain private uses of artificial intelligence in sectors such as Health, Education, Justice, Security, and Economy and Finance must adapt to these demands.
The Regulation constitutes progress because it introduces a risk-based approach, establishes transparency requirements, incorporates human oversight mechanisms and promotes technical standards for the development and responsible use of artificial intelligence. Not all systems or uses represent the same level of risk and, therefore, do not require the same controls.
The opportunity is broad given that artificial intelligence can improve public services, accelerate processes, expand the analysis capacity of organizations, increase productivity and generate new opportunities for companies and citizens. However, incorporating these tools is not just a technological decision, since ethical responsibilities must be assumed, risks evaluated, monitoring and supervising results, and establishing criteria to protect citizens when any of these systems affect their rights or living conditions.
Complying with the Regulation requires much more than approving policies and protocols, since it requires budget, computing infrastructure, processing capacity, quality data, security, technological integration, specialized personnel, monitoring and effective human supervision mechanisms. In services that serve large volumes of users, cloud, processing, storage, maintenance and updating costs generate permanent expenses.
However, the decree that approves the Regulation provides that its implementation and sustainability be financed from existing institutional budgets, without demanding additional resources from the Public Treasury. The consequence may be unequal application, capable of deepening existing gaps in access to quality public services. In this sense, organizations with greater digital maturity will be in better conditions to meet the new demands, while others may face difficulties in scaling their services, adequately supervising their systems or maintaining sufficient levels of quality and security.
Quality cannot be assessed with a single criterion either, as an acceptable level of precision for a support tool may be insufficient when a decision affects a person’s rights, access to services or economic conditions. Therefore, it is necessary to develop criteria that allow measuring reliability, error, bias and performance according to the level of risk and the purpose of each application, an aspect that still requires greater operational precision.
The strategic value of the Regulation lies in establishing conditions to safely adopt artificial intelligence, preserving innovation and competitiveness. If this balance is achieved, it can translate into better services for citizens, new opportunities for companies and a real contribution to the economic and social development of the country.
