How is AI defined under the EU AI Act?
Understanding the Definition of AI in the EU AI Act
The upcoming EU AI Act is set to impose significant regulations on the development and deployment of artificial intelligence within the European Union. A crucial first step in understanding these regulations is knowing which models are actually classified as “Artificial Intelligence” under the European legislation. Fortunately, the legislative text provides a clear definition.
Under the EU AI Act, an Artificial Intelligence 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"
Key Factor 1: Autonomy – More Than Just Automation
A primary factor in this definition is the system’s autonomy. A necessary condition for a model being classified as an AI system is that it can autonomously transform certain data inputs into outputs without direct human intervention required to generate outputs from inputs. This distinguishes AI from simple automation, where tasks are pre-programmed but lack independent decision-making.
Key Factor 2: Inference – The Ability to Learn and Adapt
The second key factor is the system’s capability to infer. Within the legislation, AI models are explicitly distinguished from simpler programming approaches like rules-based systems. While a rules-based system can autonomously transform inputs to outputs without human intervention, the decision rules themselves are exclusively designed by humans.
As stated in Recital 12 of the AI Act, “a key characteristic of AI systems is their capability to infer. This capability to infer refers to the process of obtaining the outputs, such as predictions, content, recommendations, or decisions, which can influence physical and virtual environments, and to a capability of AI systems to derive models or algorithms, or both, from inputs or data. The techniques that enable inference while building an AI system include machine learning approaches that learn from data how to achieve certain objectives, and logic- and knowledge-based approaches that infer from encoded knowledge or symbolic representation of the task to be solved.“
What’s In and What’s Out: Examples Under the EU AI Act Definition
In summary, AI models as defined under the EU AI Act exclude human-defined rules-based systems but include a wide range of data-driven approaches to decision-making:
- Machine learning approaches (supervised, semi-supervised, or unsupervised)
- Logic- and knowledge-based approaches (where the logic is learned from data, not explicitly defined by humans)
- Classical statistical approaches (where the data transformation into outputs is learned from data, such as regression models or statistical search and optimization methods)
Excluded
- Traditional, human-defined rules-based systems
Conclusion: A Broad Definition with Far-Reaching Implications
The EU AI Act ties its definition of AI models to the models’ autonomy as well as their capability to infer, taking on a broad definition of Artificial Intelligence within the landscape of different data-driven approaches. This broad definition means that many existing systems will fall under the scope of the Act, requiring careful compliance and consideration.