Amendments to EU’s proposed AI regulation – Parliament to be informed by a follow-up Union communication
The EU’s objective is to create common rules for the use of artificial intelligence (AI). The regulation aims to ensure that the use of artificial intelligence is safe for all companies and people. Amendments have been made to the original proposal and the Government will inform Parliament of the amendments.
On 26 October 2022, the Government submitted a follow-up Union communication concerning the proposed artificial intelligence regulation. When informing Parliament of EU affairs that fall within the competence of Parliament, the Government issues a Union communication.
The first Union communication on the proposed artificial intelligence regulation was submitted to Parliament in May 2021. The working group of the Council of the European Union made amendments to the proposal, which will now be presented to Parliament in the follow-up Union communication.
Definition of artificial intelligence should only cover genuine AI
The amended proposal has narrowed the definition of artificial intelligence. It is important for Finland that the definition of an AI system only covers systems that can genuinely be considered artificial intelligence.
In Finland’s view, the regulation should not apply to rules-based automation, i.e. systems that automatically apply human-made rules and instructions. Finland has taken the initiative in narrowing the definition of artificial intelligence to ensure that the use of certain automated decision-making systems can continue, for example.
Provisions concerning general purpose AI systems have been added to the proposed regulation. General-purpose AI systems are used for many different purposes, for example to identify and produce images and sound and for translation purposes.
In addition, the amended proposal includes regulation on the testing of high-risk AI systems in a real-life environment.
EU Member States aim to agree on a general approach in December
According to Finland, the regulation concerning experimentation and testing of AI systems should support innovation and not set unreasonable barriers to the market entry of systems.
It is important for Finland that the requirements set for general-purpose AI systems are proportionate. Moreover, the EU should critically assess to which extent the requirements set for high-risk AI systems should also apply to general-purpose systems. Finland’s objective is that the obligations laid down in the regulation do not apply to AI systems used privately by households.
The Czech Presidency aims to reach a general approach on the matter with the EU Member States at the Telecommunications Council meeting in December 2022.
Maria Hauptmann, Chief Specialist, Ministry of Economic Affairs and Employment, tel. +358 295 047 334
Kristine Alanko, Specialist, Ministry of Economic Affairs and Employment, tel. +358 295 047 344