The Council of the European Union (EU) on Tuesday gave the final green light to the Artificial Intelligence (AI) Act, the world's first legislation for this technology, to safeguard fundamental rights in the EU area, which will come into force within weeks.
In a statement, the European institution that brings together the EU's member states said it had "today approved an innovative law aimed at harmonizing the rules on artificial intelligence, the so-called Artificial Intelligence Act".
"The flagship legislation follows a 'risk-based' approach, which means that the greater the risk of causing harm to society, the stricter the rules. It is the first of its kind in the world and could set a global standard for the regulation of AI," underlines the Council of the EU.
Several rules are defined for classifying AI systems as high-risk, with the prohibition of certain practices, the inclusion of general-purpose AI models, a new governance architecture and measures to support innovation and transparency and the protection of fundamental rights.
The law also provides for fines for infringements that are set as a percentage of the infringing company's overall annual turnover in the previous financial year or a predetermined amount, whichever is higher, with small and medium-sized enterprises and startups being subject to proportionate administrative fines.
"The new law aims to promote the development and adoption of safe and reliable AI systems across the EU single market, by both private and public actors. At the same time, it aims to ensure respect for the fundamental rights of EU citizens and to stimulate investment and innovation in the field of artificial intelligence in Europe," justifies the Council of the EU.
The AI act applies only to areas covered by EU legislation and provides for exemptions, such as systems used exclusively for military and defense purposes, as well as for research purposes.
The European Parliament's approval in mid-March of what it called a "historic law" to regulate AI in the EU contributed to Tuesday's endorsement.
After final adoption by the member states (in the Council), the law will enter into force 20 days after its publication in the Official Journal, with an adaptation period of two years.
It will thus be fully applicable 24 months after its entry into force, except for prohibitions on prohibited practices (applicable six months after the date of entry into force), codes of conduct (nine months after entry into force), general AI rules, including governance (12 months after entry into force) and obligations for high-risk systems (36 months).
The regulation establishes a series of obligations for AI based on its potential risks and level of impact, stipulating a ban on certain applications, limitations on the use of biometric identification systems by law enforcement, exemptions applicable to law enforcement, obligations for high-risk systems and transparency requirements.
This is the first regulation aimed at AI, although the creators and developers of this technology are already subject to European legislation on fundamental rights, consumer protection and safety rules.
