Regulation on artificial intelligence has already been published

On July 12, 2024, finally published in the Official Journal of the European Union the Regulation 2024/1689 of the European Parliament and of the Council of June 13, 2024, which lays down harmonized rules on artificial intelligence (known as “AI ACT”). As stated in article 1 of the AI ACT, this regulation has four primary purposes: to improve the internal market, to promote the uptake of human-centric and trustworthy AI, and to protect health, safety, fundamental rights, democracy, rule of law, and environment, from harmful effects of AI Systems, while supporting innovation.

Providers and deployers placing AI systems or general-purpose AI models on the European market or putting them into service shall be aware of the new obligations that will be applied to them.

First, they should confirm if they are trading and using an AI system as defined by this Regulation in article 3 (1): a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.

Second, they should be aware that there will be prohibited AI practices, such as manipulative or deceptive techniques, that distort a person’s behaviour and the evaluation and classification of natural persons, leading to unfavourable treatment, as listed in Article 5 of the AI ACT.

Third, providers and deployers must confirm whether their systems are high-risk, limited-risk, or minimum-risk, as explained in detail in the AI Act policy. They can use the compliance checker to help them understand whether or not AI ACT is applicable to the system they are creating or using.

This is important because several obligations will be applied according to the level of risk. Providers and deployers of High-risk systems will have to comply with several new obligations, amongst others, pursuant to articles 9 and subsequent: transparency, human oversight, risk management system, and registration.

It has been widely discussed how these systems can become explainable and, therefore, possible for any human supervision. Authors are currently debating the issue of AI explainability. For instance, there is a discussion regarding whether Explainable AI is feasible.

Providers and deployers of General-purpose AI models will have to comply with transparency rules and adopt codes of conduct, as foreseen in articles 53 and subsequent.

Entities such as the European AI Office and the European Artificial Intelligence Board have been established to assist companies in implementing this regulation, as stated in articles 64 and subsequent. Therefore, we recommend that companies follow all their initiatives, communications, and guidelines. Please check some events hosted by the AI Office here https://digital-strategy.ec.europa.eu/en/policies/ai-office and by European Artificial Intelligence Board https://digital-strategy.ec.europa.eu/en/news/commission-hosts-high-level-meeting-upcoming-eus-ai-board-drive-ai-act-implementation-forward.

Additionally, a standardization process has been put in place to create the necessary technical norms to help developers and other professionals understand the techniques that will be executed to comply with AI ACT rules. A preliminary standardization work plan in support of AI ACT has been released and can be consulted in [5].

This regulation shall apply from 2 August 2026, with some exceptions regarding prohibited AI Practices and AI High-risk systems, provided for Article 113.

The following two years will be crucial for understanding how to implement these new rules fully. Let us remain alert.

Regulation on artificial intelligence has already been published

References

Citation