AI Act

EU states give final approval to AI Act

Among other things, the AI Act requires AI-generated content to be labeled.

(Bild: Jannatul Koraise - stock.adobe.com)

The EU member states voted unanimously in favor of regulating artificial intelligence in Brussels on Tuesday, giving their final approval to the so-called AI Act. The EU law divides AI systems into different risk categories and is intended to regulate them accordingly. AI-generated texts, images and videos will also have to be labeled in the future.

Austria's State Secretary for Digitalization Claudia Plakolm (ÖVP) welcomed the labelling requirement ahead of the meeting of EU ministers responsible for telecommunications on Tuesday, particularly with regard to young people. They are flooded on the internet "by a myriad of images and videos (...) from influencers who mostly do not even exist, who were created with just a few clicks of the mouse" and who "massively influence their own perception of beauty".

However, Plakolm also sees AI regulation as an opportunity for Austria as a business location. The country is well positioned in terms of "application research and basic research".

Certain applications such as biometric categorization based on sensitive characteristics or the untargeted reading of facial images from the internet or surveillance cameras are prohibited under the AI Act - there are exceptions for security authorities. This is a fine line, Plakolm told journalists in Brussels. It requires "clear regulations", but also "the courage to be free".