The European Parliament approves the “EU AI Act”, legislation that regulates the use of artificial intelligence in the European Union

The European Parliament approves the “EU AI Act”, legislation that regulates the use of artificial intelligence in the European Union

On March 13th, the European Parliament approved the Artificial Intelligence Act (AI Act). The new legislation, which deals with the use of artificial intelligence (AI) technologies in the European Union, will come into force within 20 days of its publication and is binding on all EU countries.

In one of the first points, the AI Act states that, by definition, AI is “a machine-based system designed to operate with varying levels of autonomy, which may have the capacity to adapt after deployment and which, for explicit or implicit purposes, infers, from the input it receives, how to generate results, such as predictions, content, recommendations or decisions that can influence physical or virtual environments”. The article also contains other important definitions such as: provider, deployer, identification and biometric verification.

Next, the legislation sets out the prohibited practices related to AI. These include: the prohibition of using AI to exploit vulnerabilities related to the age, disability or socioeconomic status of specific individuals or groups; the indiscriminate collection of facial images from the internet or closed-circuit television systems to create or expand facial recognition databases; and the prohibition of real-time biometric identification by police authorities in public spaces, although there are exceptions.

The AI Act also adopted a system that classifies high-risk AIs. In the legislation, it is defined that an AI is only not high-risk when “there is no significant risk of harm to the health, safety or fundamental rights of natural persons, including not materially influencing the outcome of decision-making”.

Regarding the development of new AIs, the regulation provides for the possibility of testing high-risk AI systems in “real-world” conditions, beyond regulatory sandboxes, as long as they have the informed consent of the data subjects involved.

Finally, regarding transparency, the act establishes that AI systems that interact directly with people must clearly state that they are AI, in addition to the obligation to provide technical information, comply with EU copyright laws and disclose information about the content used in training.

The text of the new legislation can be accessed via the link: EU Artificial Intelligence Act