Is the EU AI Act a Jenga piece that can easily be removed from the regulatory tower?

Here are some key points from the “AI Regulation – a critical comment” workshop at the Alpine Privacy Days Conference, courtesy of Florent Thouvenin, Professor of Information and Communications Law, The Faculty of Law of the University of Zurich, Switzerland

  • Would it have been better to just not define AI system in the AI Act?
  • Some practices are prohibited when using AI, but they are not necessarily prohibited from doing the same thing without using AI.
  • There is the AI Act, Standards and generally applicable laws and sectoral laws. We could arguably get to the same result if we remove the AI Act from the picture.
  • Transparency is very important, but is the focus on AI technology the right way? What about a human using photoshop? What do we do when most content will be AI generated?
  • Is the AI Act a version of the “Red Flag” law, where a person had to walk in front of cars in order to ensure that they don’t cause damage?
  • Is it better to wait and see where the tech us going? To better ensure innovation is hindered due to this? This is the approach in Switzerland, to wait and rely on existing laws.