EU Artificial Intelligence Act: what happens now?

  Articoli (Articles)
  Jacopo Biagi
  29 June 2024
  3 minutes, 31 seconds

Translated by Alessandra Fumagalli 

In the last few years, the entire world, starting from the experts to the amateurs, started to know about Artificial Intelligence (AI) and its numerous applications. This technology, apparently prodigious and extremely versatile, raised some worries speaking of privacy issues, data security and its uses. 

In order to define some clearer guidelines and guarantee users’ security, the European Union presented the AI Act, a regulation that wants to regulate the uses and developments of artificial intelligence. On 13rd, March, the European Parliament was called to vote on the bill, which was largely approved with 523 favorable votes, 49 dissenting votes and 49 abstensions. 

Introduction to AI Act

This Act is the first and most complete regulation on Artificial Intelligence. The key point of the law is the classification of AI based on some risk categories. The highest unacceptable risk deals with those artificial intelligence systems that are a threat for some people. Social classification systems, biometrical identification and systems able to manipulate people’s behavior are considered unacceptable risks, therefore their use is forbidden in Europe. 

As far as high risk systems are concerned, they are regulated in the Act, which imposes, to suppliers and developers, heavy transparency and security’s duties for their use. 

A small section of the Act deals with those AI’s projects that have a limited risk for the users; in these terms, there are lighter transparency’s duties for the suppliers and it must be granted that final users will be aware of the fact that they are interacting with an AI system. 

For minimum risk systems, which are already in the European markets such as videogrames and some software for PC, there aren’t any particular duties to the developers. 



Reactions to the law

The news about the recent approbation of a law that regulates artificial intelligence causes opposing reactions in the industrial panorama and among European countries. Sectors that work with AI accepted this law, because it involves a clear and definitive regulation of a sector that wasn’t regulated. Some giants of technology, like Microsoft and Google, and other experts have highlighted the need for clear and universal standards for a secure and ethical development of this technology. 

However, critics and worries for the final law came. The major doubts come from small companies and startups which operate in the technological sector and develop artificial intelligence projects. Smaller companies, indeed, are at risk of being endangered and penalized by high-risk developers’ strict duties. Small companies use a small quantity of capital and are not able to economically satisfy the AI Act’s duties and are disadvantaged than other companies, like the Big Tech. Due to a strict law, whose aim was to protect citizens, there will be a situation where bigger companies would dominate the market and research. Max von Thun, director for Open Markets Europe, stated that artificial intelligence can “consolidate the extreme power that few technological companies already have in our personal lives, economies and democracies”.  

However, despite the high worries of a direct favoritism towards big tech, some startups approve the new law. The co-founder and CEO of Giskard-AI, Alex Combessie, describes it as an “historical and relief moment”. Despite some adjustments to the new duties causing added work to small agencies, Combessie is confident about implementation of controls in the small companies.  

The European Union, with this law, took an important step towards the technology field: with clear and esaustive rules, it wants to become competitive and attractive to new investors, competing with the American and Chinese markets which are cutting-edge in the technological sector. The challenge will be to find a new balance able to protect citizens and promote innovation and research of AI systems which can prosper the common European market. 

Mondo Internazionale APS - Riproduzione Riservata ® 2024

Share the post

L'Autore

Jacopo Biagi

Categories

Ambiente e Sviluppo

Tag

artificial intelligence Ai Act European Union technological development europe AI european law