Agreement reached on EU’s regulation on artificial intelligence
Agreement was reached at the end of 2023 on the European Union's (EU) regulation on artificial intelligence (AI). The AI Act is expected to enter into force later this year, once the legislative text has been finalised and the regulation has been formally adopted. What will the AI Act mean in practice?
EU’s AI Act aims to ensure that AI is used in the EU in a safe way that respects fundamental rights and EU values. Through sensible and sufficiently flexible rules, EU strives to be the first in the world to ensure that citizens can trust systems that make use of AI.
Risk-based regulation on AI
Towards the end of the negotiations, many items on the European Parliament’s list of prohibited AI applications were moved to a high-risk category. To counterbalance the high-risk category, an assessment and authorisation framework was created to help reduce risks.
The definition of high-risk became problematic during the negotiations. Ambiguity and incomplete regulations have proven to be a barrier to the use of AI, with investments not being made for fear of significant sanctions.[1] According to a report by the Ministry of Economic Affairs and Employment of Finland and Demos Helsinki, the use of the AI Act requires clear regulation, communication and capacity building as part of broader digital regulation.
For Finland, the AI Act does not seem to pose insurmountable problems. Finland has been able to influence the content of the Act, for example, artificial rule-based automation is no longer covered by the Act. However, it is difficult to define the limits.
It is likely that the earnings-related pension field and social security will be included in the high-risk category and its obligations, so the development and use of AI will have to be applied within these limits. The final wording will be known when the final text of the law is distributed to stakeholders for reading.
When will the AI Act enter into force and what happens next?
A political agreement on EU’s AI Act was reached by the Member States, the European Commission and the European Parliament in December 2023, with a phased implementation period of two years from entry into force. Prohibited uses of AI will come into force as early as six months after the Act enters into force. The Act is tentatively expected to enter into force before the European elections in early June 2024.
According to a Commission press release, scientific research and development activities will be excluded from the scope of the act. Evaluation models and transparency obligations will be established for the use of generative AI. In addition, the Act would make testbeds for AI mandatory.
The Act will be directly applicable legislation in Member States, but will require national preparations for implementation. The Finnish Centre for Pensions is actively involved in the national implementation work.
Read more
Legal framework on artificial intelligence: AI Act – Shaping Europe’s digital future (europe.eu)
Artificial intelligence act: Council and Parliament strike a deal on the first rules for AI in the world (europe.eu)
[1]Graduated sanctions: €35 million or 7% of the previous financial year’s turnover for breaches of prohibited AI applications, €15 million or 3% of turnover for breaches of obligations under the AI Act and €7.5 million or 1.5% of turnover for providing incorrect information.