In previous articles we had mentioned that the European Parliament was expecting to approve the historic milestone of the Artificial Intelligence Law , and this has happened.
On Wednesday, March 13, 2024, the European Parliament approved the legislation that will regulate various aspects of Artificial Intelligence, being supported with 523 votes in favor, 46 against and 49 abstentions.
The aim of this Act is to protect fundamental rights, while at the same time promoting innovation and making the European Union a leader and a vanguard in this sector.
In this regard, the Regulation provides for a series of prohibited applications , such as applications that, through the use of Artificial Intelligence, violate the rights of citizens, highlighting biometric systems or images and recordings to create facial recognition databases.
Additionally, the recognition of emotions is prohibited in workplaces or educational environments, among others.
However, in the case of biometric identification systems, they will also be prohibited a priori by the security forces, although there are exceptions .
In the event that the risk is high , in addition to the usual obligations, companies must comply with a series of additional requirements, in addition to having systems that evaluate and reduce risks, being transparent and precise when providing data.
This transparency is one of the basic principles in the use of Artificial Intelligence, which must respect Copyright and publish detailed summaries of the content used.
Claims may be filed against companies that do not comply with the requirements. And, as an example: Influence on elections. If the conditions set out in the Regulation are not met, the regulations themselves provide for fines for non-compliance that may reach the amount of thirty-five million euros.
As for SMEs and Emerging Companies , States that subscribe to the Artificial Intelligence Law must facilitate measures to support innovation so that they can incorporate AI.
For its part, ChatGPT from OpenAI or Bard from Google, has also been the subject of discussion, as generative artificial intelligence, reiterating the obligation of this type of applications to comply with the principle of transparency and, with the corresponding Copyright , excuse the redundancy, to its authors.
Although the regulations have been approved, they are subject to legal-linguistic verification and final approval through the error correction procedure before they can be formally adopted by Congress.
Furthermore, it must be published in the Official Journal, entering into force twenty days later and being fully applicable twenty-four months after its entry into force, with exceptions:
- Bans: 6 months later.
- Codes of good practice: 9 months later.
- General use and governance: 12 months later.
- Obligations in very high risk systems: 36 months later.
Despite the above, it is expected to come into full force in mid-2026.
