Some Artificial Intelligence (AI) technologies have been around for decades, but advances in the availability of large amounts of data and new algorithms have boosted the use of AI models in recent years. Some AI systems create risks that need to be addressed, and there has been an exponential growth in regulation in this regard.

Artificial Intelligence: regulatory landscape

Watch video

Executive summary

The growing use of AI models in all sectors has led to an exponential increase in regulation in recent years. The European AI Act and the US Bill of Rights on AI stand out as reference regulations. Moreover, the regulation of this matter is also growing in other countries around the world, such as Canada, Brazil, Japan and China.

Main Content

This Technical Note aims to provide an updated overview of the regulatory landscape and the international standards that make up the regulatory framework and best practices for Artificial Intelligence. The following should be highlighted:

  • At global level, the Principles for AI were published in 2019 by the OECD. This document sets out principles for responsible and trustworthy AI governance. It also sets out recommendations for the integration of AI into national policies and encourages international governmental cooperation for safe AI.
  • At European level, the draft AI Act was published by the European Commission in 2021. This document proposes a classification of AI practices into the following levels: i) prohibited practices; ii) high-risk AI systems; and iii) low-risk AI systems. In addition, it includes transparency requirements for systems that: i) interact with humans; ii) are used to detect emotions; or iii) generate or manipulate content.
  • In the USA, the Blueprint for an AI Bill of Rights sets out five AI principles to guide the design, use and deployment of automated systems to protect the rights of US citizens. The principles are: i) safe and effective systems; ii) algorithmic discrimination protections; iii) data privacy; iv) notice and explanation; v) Human alternatives, consideration and fallback.
  • In other geographies, some publications were also released. In Canada, the Law on IA and Data (AIDA) was published in 2022. In Brazil, the first draft Regulation on IA was published in the same year, and in Japan, the Governance Guidelines for the Implementation of IA Principles were also published in 2022. Other countries like Italy have implemented measures for the management of generative AI services, including a ban on the use of ChatGPT that lasted a few months.

Download the technical note on Artificial Intelligence: regulatory landscape.