EU leaders reached agreement on a crucial AI law after 36 hours of talks

Members of the European Parliament, representatives of member states, and experts from the European Commission have reached a political agreement on a law on artificial intelligence (AI), which is the EU’s first attempt to regulate the new technology.

After more than 36 hours of negotiations lasting three days, representatives of the three EU institutions managed to reach an agreement on such controversial issues as predictive applications, face recognition, and the use of artificial intelligence by law enforcement agencies, Politico reported.

Experts anticipate that further technical revisions will be necessary to enhance the political agreement that has been achieved.

“The EU becomes the very first continent to set clear rules for the use of AI,” Internal Market Commissioner Thierry Breton posted on his account on X. “The [AI Act] is much more than a rulebook — it’s a launchpad for EU startups and researchers to lead the global AI race.”

“The AI law is much more than a set of rules; it is a launching pad for startups and researchers from the EU to lead the global AI race,” the EU commissioner added.

Lawmaker Dragoș Tudorache, one of the AI Act’s co-rapporteurs in the European Parliament, wrote, “We are the first in the world to set in place real regulation for [AI], and for the future digital world driven by AI, guiding the development and evolution of this technology in a human-centric direction.”

According to AFP, the launch of ChatGPT, a text generator from California-based OpenAI that can write essays, poems, or translations in seconds, had a decisive impact on the start of the negotiation process.

This system, as well as systems capable of creating sounds or images, has opened up the world to the enormous potential of artificial intelligence but also to certain risks. For example, the spread of fake photos on social media has shown the danger of manipulating public opinion.

Systems considered “high-risk” and used mainly in sensitive areas such as critical infrastructure, education, human resources, law enforcement, etc., are subject to a list of rules in the EU project. These systems will be subject to a number of obligations, such as ensuring human control over the machine, drafting technical documentation, or implementing a risk management system.

The legislation provides for the introduction of special supervision over artificial intelligence systems that interact with humans. It will force them to inform the user that they are interacting with a machine.

The document also provides for the rare use of prohibitions, which will apply mainly to those applications that contradict European values. An example that should not be imitated is China, which has systems for assessing citizens, conducting mass surveillance, and implementing remote biometric identification of people in public places.

However, EU countries have been granted freedom to engage in certain types of law enforcement activities, such as combating terrorism.

Additionally, the agreement stipulates that the European Commission will establish a European Office for Artificial Intelligence, and European legislation will be equipped with surveillance measures and sanctions. For the most serious offenses, the agreement will impose fines of up to 7 percent of turnover, with a lower limit of 35 million euros.

Germany, France, and Italy have reached an agreement on how to regulate artificial intelligence in the future.

Before that, the National Cyber Security Centre of the UK stated that artificial intelligence poses a threat to the next parliamentary elections in the country, and cyberattacks by hostile countries and their proxies are becoming increasingly difficult to track.

Signing the first international declaration dedicated to this fast-growing technology, the UK, the US, the EU, Australia, and China have all agreed that artificial intelligence poses a potentially catastrophic threat to humanity.