Since its emergence and faced with the rapid pace of its evolution, artificial intelligence (AI) has raised concerns on the old continent. The European Union (EU) has taken the lead in pushing for regulations whose objective is to curb the potentially negative effects that the development of this technology could entail.
European regulation on AI unanimously approved
European Union (EU) member states have unanimously approved the first comprehensive regulation on artificial intelligence (AI). This is an important step forward in regulating the potential harmful effects of AI.
The political agreement, concluded in December 2023, was subject to technical adjustments for more than a month. Work, concluded with the presentation on January 24 of the final version of the law on AI, by the Belgian presidency.
Despite initial reluctance and resistance from some EU heavyweights such as France, Germany and Italy, a compromise was finally found. This, after a 36-hour marathon of negotiations.
The compromise resulted in a phased approach, imposing horizontal transparency rules for all AI models. Additional obligations for those considered to present systemic risks are also decided.
The road will have been long
The main point of contention was the regulation of powerful AI models. France and other countries were pleading for lighter rules to support European startups. Meanwhile, the European Parliament was pushing for stricter rules.
Finally, France, initially reluctant, agreed to support the text subject to strict conditions. The latter must guarantee the development of competitive AI models and promote a balance between transparency and protection of commercial secrets. This is done by avoiding overloading companies with high-risk bonds. Still, countries like Slovakia and Austria have expressed other concerns. These relate to critical terms, international alignment and the impact on data protection and consumer rights.
In any case, the European Commission is preparing for the establishment of the European Office for Artificial Intelligence, which is currently being created. The next steps include adoption by European Parliament committees and a plenary vote, followed by formal approval at ministerial level.
The European regulation on AI will come into force 20 days after its publication. Bans on prohibited practices will take effect after six months and obligations relating to AI models will begin to apply after one year. Full implementation of the rules will take place within two years, with the exception of provisions concerning high-risk AI systems, which are subject to an additional one-year deadline.
Maximize your Tremplin.io experience with our ‘Read to Earn’ program! For every article you read, earn points and access exclusive rewards. Sign up now and start earning benefits.
