Artificial Intelligence, OECD principles and recommendations (Part I)

The document presents guidelines, five principles and five recommendations, aiming to set international standards able to guarantee that AI systems.
The Organization for Economic Cooperation and Development (OECD) is an international organization dedicated to the design of global policies to improve people’s economic and social well-being globally. Considering, therefore, that Artificial Intelligence (AI) has implications for the world, social and economic transformations, as well as holds the potential to contribute positively to economic activity, innovation and productivity. The OECD has developed “Recommendation of the Council on Artificial Intelligence” in order for members and adherents to promote and implement their content to promote public trust and benefits to all interested parties.

The document presents guidelines, five principles and five recommendations, aiming to set international standards capable of guaranteeing that AI systems, in all phases of their life cycles, will be robust, safe, fair and reliable. The primary purpose is to guide governments, organizations and other actors in the design and implementation of AI systems, ensuring centrality in the human person, together with an adequate responsibility regime. In this way, it is expected that the benefits provided by this new technology will be widely available to society.

Internationally, more than four dozen countries have already adhered to these guidelines. In addition, the G20, in June 2019, adopted principles for human-centered AI, based on the document prepared by the OECD, recognizing the need to achieve common goals for global economic growth.

In Brazil, the public consultation opened in early 2020 by the Ministry of Science and Technology reported that the structuring of AI strategy in the country would go through the adoption of the OECD guidelines. Bill 21/2020, likewise, explicitly mentions the OECD guidelines for the development of this new technology. The justification explains that Brazil’s adherence to the document in May 2019 makes, according to its author, “appropriate the edition of legislation on the matter”.

The complementary principles applicable to all stakeholders are: (1) inclusive growth, sustainable development and well-being; (2) values ​​centered on human beings and equity; (3) transparency and explanability; (4) robustness, security and protection; (5) “accountability”. In line with these principles, the OECD has made five recommendations to member countries and those that adhere to the document. These are focused on national policies and international cooperation: (1) investing in AI research and development; (2) promoting digital ecosystem for AI; (3) organize a political environment favorable to AI; (4) strengthening human capacity and preparing people for the transformation of the labor market; (5) international cooperation for reliable AI.

Although principles and recommendations do not legally bind OECD member countries and adherents to the document, they provide perspectives on national laws and regulations that are under discussion or drafting. The relevance, therefore, of addressing this grouping in detail is associated with the fact that these guidelines become international standards, that is, a reference for all interested parties.

By: Wilson Sales Belchior