Last week, EU ministers unanimously gave their final approval to the Artificial Intelligence Act (AI Act), a landmark new law that regulates the use of the transformative technology in “high-risk” situations, such as law enforcement and employment.

“The adoption of the AI Act is a significant milestone for the European Union. This landmark law, the first of its kind in the world, addresses a global technological challenge that also creates opportunities for our societies and economies,” said Belgian secretary of state for Digitisation, Mathieu Michel

At a global level, AI is also being closely watched. Last week, more than a dozen countries and some of the world’s biggest tech firms gathered in Seoul for an AI summit co-hosted by the United Kingdom and South Korea. They pledged to cooperate against the potential dangers of AI, including its ability to dodge human control.

The EU’s regulation allows or prohibits the use of AI depending on the risk it generates for people and identifies high-risk systems that can only be used if it is demonstrated that they respect fundamental rights.

Which AI is being targeted by the law?

AI systems for biometric categorisation based on political, religious, philosophical beliefs or race and sexual orientation will be prohibited. Systems that rate people based on their behaviour or personal characteristics or AI capable of manipulating human behaviour will be banned from use. 

Systems to expand or create facial databases captured indiscriminately through the internet or audiovisual recordings will also be prohibited. However, the regulation allows exceptions, so that security forces may use biometric identification cameras, always with judicial authorisation, for example to prevent a terrorist threat.

Furthermore, the law stipulates that content generated with AI, such as AI-generated texts, images or videos, will have to be labelled as such. This should help protect viewers against misleading content, such as deepfakes.

High-risk systems will have to obtain certification from approved bodies before they can be put on the EU market. A new “AI Office”’ will oversee enforcement at EU level. Failure to comply with the regulation can lead to fines up to 35 million Euro or seven percent of a company’s annual revenue, depending on the type of offender.

Portuguese push to tackle AI

The European Commission proposed the first draft of the AI Act in April 2021. At the time, Portugal held the rotating Presidency of the Council of the European Union. 

One of the ambitions of the Portuguese government – which had a socialist majority at the time – was to push for the adoption of the first EU law on AI, based on transparency and respect for users’ rights.

“We attach great importance to the legal framework for artificial intelligence. Right now, it’s clear that artificial intelligence is the basis for enhanced productivity and has great potential for growth,” said the then Minister of Economy, Pedro Siza Vieira, in January 2021.

The AI Act already received the backing of a majority of Members of the European Parliament in March 2024. The two rapporteurs for the file were Italian MEP Brando Benifei and Romanian MEP Dragoș Tudorache.

Following Tuesday’s unanimous vote among ministers to finalise the AI Act, it must be signed by the presidents of the EU legislature and then published in the Official Journal of the EU. It then technically becomes law 20 days later, but its various provisions will come into effect gradually over the following two years.

Challenges for the implementation phase

With the legislative process for the AI Act at an end, EU member states will have to implement the regulation into their respective national legislation – with some representatives at last week’s meeting pointing out future challenges.

“We tried to find a balance between the two approaches so that artificial intelligence develops and so that, for example, smaller companies are not overwhelmed by bureaucracy, but at the same time so that the use of artificial intelligence in the European environment has its barriers,” said Czech Deputy Prime Minister for Digitisation and Minister of Regional Development Ivan Bartoš, who represented the country during the meeting of EU ministers.

The focus on the implementation process was echoed by Ivan Ivančin, state secretary of the Ministry of Investment, Regional Development and Informatics of the Slovak Republic, who said “it represents an opportunity for learning”. He added that “at the same time, it is important to implement concrete steps that will create a solid foundation for further development”.

Bulgaria’s caretaker government supported the adoption of the AI Act. Bulgarian Entrepreneurial Association CEO Dobromir Ivanov noted that the AI Act is a step in the right direction, but stressed that it is important to watch what will be done from now on. He said that implementation should not be too restrictive or destroy local companies. 

During a discussion at the conference “ygeiamou 2024 – The Health System and the Challenges of the Future” in Athens, Greek Prime Minister Kyriakos Mitsotakis spoke about the application of AI in healthcare. In particular, he pointed out that when processing patient data with AI, for example, it is important to manage it securely.

AI experts wanted

Slovenian Minister for Digital Transformation, Emilija Stojmenova Duh, warned of the lack of AI experts. “At the moment, not only in Slovenia, but in Europe in general, we do not have enough qualified experts, so we need to work on that.”

She thinks experts from all member states should participate in the EU regulatory authority for AI, which has already been set up by the European Commission. This will also make it easier to implement the act in member states, she said.

In Romania, the construction of the first research institute on AI began last week at the Technical University of Cluj-Napoca. The institute will focus on developing AI solutions for vital areas such as health, transport, cyber security and more, and will offer students and researchers cutting-edge resources and the chance to work on projects with global impact.

Fact check: Fake news video created with artificial intelligence

With the use of artificial intelligence, anyone can create artificial images or videos, which are often used to mislead viewers of such content – so-called deepfakes. In a video circulating online in February and March 2024, a news reporter is seen saying that French President Emmanuel Macron had to cancel his trip to Ukraine to meet President Volodymyr Zelensky because of an alleged plot to assassinate him.

The fact-checking team by German Press Agency dpa explained that the video was fake and that both the audio – the presenter’s voice – and the video footage had been altered using AI. The video circulated in several languages, including German.
Read the full fact check here:

This article is published weekly. The content is based on news by agencies participating in the enr.