European Union published Regulation 2024/1689 (“AI Act”) on 12 July 2024 after three years the proposed text was released by the European Commission in April 2021.(European Commission, COM/2021/206.) The AI Act is considered to be world's first comprehensive AI regulation and it is a broad regulatory framework designed to establish the basis for AI governance across the EU, with significant global influence on companies and future legislation.
Under the Act an “AI System” is defined as “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.” Following the definition, AI Systems are classified in 4 categories based on a risk-based approach; (i) Unacceptable Risk, (ii) High Risk, (iii) Limited Risk and (iv) Minimal Risk. The assessment of risk level that an AI System possesses is identified by factors such as its interference with human rights and changes that it can make in the employment level. This classification then allows the AI Act to impose certain responsibilities on AI systems according to their level of risk. Consequently, while AI systems that falls into unacceptable risk category are strictly prohibited, other levels have decreasing amounts of obligations and compliance requirements.
In addition to the risk classifications, the AI Act also introduces the term general-purpose AI technologies and divides this concept into two different definitions:
- General-purpose AI models (GPAIM): “an AI model, including where such an AI model is trained with a large amount of data using self-supervision at scale, that displays significant generality and is capable of competently performing a wide range of distinct tasks regardless of the way the model is placed on the market and that can be integrated into a variety of downstream systems or applications, except AI models that are used for research, development or prototyping activities before they are placed on the market.”
- General-purpose AI Systems (GPAIS): “an AI system which is based on a general-purpose AI model and which has the capability to serve a variety of purposes, both for direct use as well as for integration in other AI system.”
If an undertaking is involved in the creation, provision or deployment of GPAIM or GPAIS they will be required to comply with several other obligations since these technologies are able to serve a vast number of purposes and tasks rather than a narrow scope.
European Union also presents a timeline for companies to adapt to the new regulatory framework. According to the timeline:
- 1 August 2024: The AI Act enters into force.
- 2 February 2025: All AI Systems that present Unacceptable Risk are banned.
- 2 August 2025: All the transparency requirements and GPAI obligations must be met.
- 2 August 2026: Obligations for AI Systems that present High Risk must be met.
- 2 August 2027: All provided AI Systems must comply with sector-specific AI regulations.
(European Union – 12 July 2024)
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.