Regulation EU 2024/1689 — better known as the AI Act — officially entered into force on 1 August 2024, with a phased application running until 2026. It is the first normative instrument in the world to comprehensively regulate artificial intelligence, and its impact will be profound for both businesses and citizens.
A Risk-Based Classification
The heart of the AI Act is its risk-based structure. AI systems are classified into four categories: unacceptable risk (prohibited), high risk (subject to stringent obligations), limited risk (transparency obligations) and minimal risk (no specific obligations).
Among the prohibited systems are, for example, social scoring systems operated by public authorities, and technologies involving subliminal manipulation. This reflects the European Union's founding values of dignity and individual autonomy.
"The AI Act is not merely a technical text. It is a declaration of values: Europe chooses to regulate before it is too late."
What Changes for the Ordinary Citizen?
For citizens, the most tangible changes concern transparency. When interacting with an AI system — a chatbot, a recommendation engine, a recruitment tool — individuals will have the right to know that artificial intelligence is being used, and to receive comprehensible explanations about automated decisions that affect them.
Particularly significant is the regime applicable to high-risk systems in sectors such as education, employment, public services and access to credit. In these areas, citizen protections are at their highest: mandatory human oversight, activity logging, and conformity assessments.
A Critical Reflection
The AI Act is an important step, but not without tensions. The main one is between innovation and protection: the obligations for high-risk systems are burdensome, and there is a real risk that European SMEs will be disadvantaged compared to large non-EU players. The Commission will need to carefully balance these interests during implementation.
As a law student specialising in EU law, I find it fascinating to observe how the European legislature is attempting — for the first time — to build a legal framework for technologies that evolve faster than the legislative cycle. The result is imperfect, but necessary.