AI ACT EU

On the 1st of August 2024 the EU Artificial Intelligence Act came into force. It is focused on clarifying wich are the main obligations and guidelines that AI must respect, safeguarding citizens well-being and their fundamental rights.

 

The AI ACT provides the defintion of “AI”, aswell as divides AI by levels of risk (“risk-based approach” according to the EU institutions). This division is made up of 4 risk levels (Minimal Risk, Specific Transparency Risk, High Risk and Unacceptable Risk) whereby the rules applied depend on the level of risk in which they are inserted. We would like to emphasise:

• Unacceptable Risk: AI Systems that imply anything related with “Social scoring”, cognitive behavioural manipulation of people or specific vulnerable groups and biometric identification and categorisation of people, shall be prohibited. However there are a few exceptions related on law enforcement matter.
• High Risk: AI Systems that use a huge ammout of citizens personal data (for exemple medical or recrutiment software), must follow very strict rules.

 

It is given a huge highlight to system data transparency and copyright law requirements. Generative AI systems, must provide that the content created was AI generated or modified. In other hand, AI systems will have to comply with avoiding from creating illegal substance. This law also foresess fines for offences measured according to the percentage of the offending company's overall annual turnover in the previous financial year or a predetermined amount.

 

The AI Act will includes an adaptation period of 2 years, so institutions and companies can gradually adapt to this new reality

 

Disclaimer: This document is intended to be distributed among Clients and Colleagues, and the information contained has a general and abstract nature since it does not spare with legal advice for the resolution of issues applied to the specific case. This information cannot be reproduced without the full consent of Global Lawyers.

GO BACK

Latest Alerts