MEPs finalize proposals for EU-wide regulation of artificial intelligence. Politicians across the spectrum want to see a ban on real-time software recognition of people or the evaluation of people’s emotions by computer technology. If the MPs’ demands are supported by individual governments, Europe will have the toughest rules for the use of artificial intelligence in the world to date.
The European Commission proposed the first ever EU law on artificial intelligence (AI) two years ago, and last December, during the Czech EU presidency, EU member states also approved their position. The ministerial debate was chaired by the Deputy Prime Minister for Digitization Ivan Bartoš (Pirates) and described the result as a “fragile balance” between innovation and the protection of the basic rights and privacy of citizens.
Among other things, Member States have supported programs to be able to recognize people remotely based on biometric data in limited situations. The ministers said that this could lead, for example, to the early identification of a dangerous terrorist.
But MEPs want to go further in the regulation of artificial intelligence. Negotiators propose that biometric recognition of people on the street can only be done retrospectively and on the basis of a court order.
“It is a consideration of two risks: either it will take a little longer to catch a terrorist, or you will collect such a large amount of biometric data on individuals that it will be possible to recognize an oppositionist in a crowd from a distance and immediately arrest him, as is already happening in Russia,” he explained to the online daily Aktuálně.cz and Hospodářské noviny MEP Marcel Kolaja, one of the shadow correspondents of the proposal.
“Most of my colleagues, including myself, support the prohibition of such identification of persons,” added the representative of the Pirates. A similar arrangement allows, for example, in the USA to use artificial intelligence for identifying people in real time.
European parliamentarians are also leaning towards banning the use of software to “read” human emotions, except in exceptional cases, for example in the healthcare sector. MEPs point out that the prospect of savings could motivate companies and institutions to use AI to evaluate candidates in tenders. According to them, it would lead to a large error rate and discrimination of people, because not everyone has a standard facial expression – the person who frowns is not necessarily angry and so on.
Stealing or just learning?
The conflicting theme then emerges with artificially generated photos and images. Getty Images is currently suing the maker of the Stability AI program in Britain for using data from millions of Getty Images photos to develop its software without the agency’s permission. Part of the MEPs therefore wants the legislation under discussion to deal with the use of foreign works by artificial intelligence. Others are against it.
“It would hinder development and innovation if copyright were to be charged here. In this case, AI does not continue to sell photos, it just learns from them,” Kolaja claims.
A number of experts are calling for the regulation of artificial intelligence in Europe. They point out that development is moving forward faster than expected, and the existing “playground” is lagging behind. For example, the Belgian and French media wrote about several schools where it turned out that their students had their particularly excellent essays written by ChatGPT-type programs. The new legislation should also cover them.
ChatGPT is probably the most popular chatbot in the world, which has been free since the end of last November and quickly gained popularity among the public. Based on simple questions, it can generate articles, essays, jokes, and even poetry. It was banned in Italy in April this year.
Pioneering technology without rules
MEPs will probably agree on their idea of artificial intelligence regulation this week. This would then be followed by negotiations on final similar standards with the member states. Both aim to approve it by early 2024.
“We are dealing with pioneering technology and using it means introducing clear rules as soon as possible,” the Financial Times quoted one of the reporters, MEP Dragos Tudorache, as saying.
Video: Artificial intelligence in art opens up a new world, says Tyl (16/08/2023)
“Progress in certain areas of artificial intelligence is enormous, it can be abused in propaganda, but I would not be afraid of technology as such,” says Jan Tyl. | Video: Daniela Písařovicová