World's first AI regulatory law is coming into effect

According to the Financial Times, late on December 8 (local time), European Union (EU) lawmakers agreed on the main provisions of the bloc's artificial intelligence (AI) management bill.

This is a major step forward in issuing the world's first comprehensive set of rules relating to the development of this type of technology.

The new law will take a two-tier approach, with higher-risk AI models facing stricter regulation, said EU Internal Market Commissioner Thierry Breton.

Specifically, general-purpose AI models such as ChatGPT or AI applications that generate images, sounds, etc. need to meet transparency requirements.

Meanwhile, the use of some powerful AI models that could have systemic impacts would be banned altogether.

This includes prohibiting the use of AI for facial recognition technology or other forms of biometric identification, except in certain law enforcement cases specified in the law.

Using AI for the government to assess an individual's reputation, commonly known as "social scoring", is also banned.

AI should also not be used to exploit vulnerable groups such as the young, the elderly, people with disabilities or people in difficult economic circumstances.

Businesses that violate the provisions of this law could face fines of up to 35 million euros or 7% of their international turnover.

"The EU will become the first continent to introduce clear rules on the use of AI. The AI ​​Act is more than just a set of rules. It is a launch pad for EU startups and researchers to lead the global AI race," Breton said.

EU lawmakers agreed on the terms after a series of discussions lasting 37 hours between representatives of the European Commission, the European Council and the European Parliament.

France, Germany and Italy have represented the European Council in asking for changes to some provisions to ease the law. Meanwhile, representatives of the European Parliament have strongly opposed this proposal.

However, the parties reached an agreement after balancing factors regarding safety and the ability to develop and apply AI.

Following the latest milestone, the AI ​​bill will undergo minor revisions. It will then be sent to member state representatives for review and approval.