Artificial Intelligence (AI), one of the most significant and rapidly evolving technological developments of recent years, is being regulated at the European Union level for the first time.
The European Parliament and the Council of the European Union have reached an agreement to establish a comprehensive legal framework governing Artificial Intelligence, as well as to amend and adapt the relevant existing legislation.
This legal framework will adopt a human-centred approach, placing particular emphasis on ensuring people’s safety when using Artificial Intelligence. Its objective is to ensure that AI systems are used safely, lawfully, and in accordance with fundamental human rights.
The main objectives of this regulatory framework are to:
- Ensure that AI systems placed on and used within the European market are safe and respect fundamental rights, freedoms, and the values of the European Union;
- Provide legal certainty to encourage investment and innovation in Artificial Intelligence;
- Improve governance and ensure the effective enforcement of existing legislation concerning fundamental rights and safety requirements applicable to AI systems;
- Facilitate the development of a lawful, safe, and trustworthy AI market while preventing market fragmentation.
AI SYSTEMS WILL BE CLASSIFIED ACCORDING TO THEIR RISK LEVEL
Unacceptable Risk
This category includes AI systems considered to pose an unacceptable threat to individuals and which will therefore be prohibited.
Such systems include:
- Cognitive behavioural manipulation targeting individuals or vulnerable groups, such as voice-activated toys that encourage dangerous behaviour in children;
- Social scoring, involving the classification of individuals based on their behaviour, socioeconomic status, or other personal characteristics;
- Real-time remote biometric identification, including facial recognition systems;
- Manipulation of human behaviour intended to circumvent or undermine an individual’s free will.
Exceptionally, certain systems may be used for law enforcement purposes, subject to prior judicial authorisation and specific legal conditions, including targeted searches for victims, the prevention of terrorist attacks, the identification and location of terrorism suspects, and investigations involving trafficking, murder, and other serious crimes.
High Risk
This category includes AI systems that may affect safety and fundamental rights. These systems are divided into two main subcategories:
1. AI systems used in products governed by EU product safety legislation, including:
- Toys;
- Aviation;
- Medical devices;
- Motor vehicles;
- Lifts and other regulated products.
2. AI systems operating in specified areas that are subject to registration requirements in an EU database, including:
- Biometric identification and categorisation of natural persons;
- Management and operation of critical infrastructure;
- Education and vocational training;
- Employment and worker management;
- Access to and use of essential private services and public services and benefits;
- Law enforcement;
- Migration, asylum, and border control;
- Assistance in the interpretation and application of the law.
AI systems classified as high-risk will be subject to assessment before being placed on the market and continuous monitoring throughout their operation and use.
Individuals affected by these systems will have the opportunity to submit complaints and obtain relevant decisions and explanations concerning their complaints or requests.
Limited Risk
AI systems falling within this category will be subject to minimum transparency requirements.
Users must be provided with sufficient information to make informed decisions, including whether they wish to continue using a particular AI system.
This category primarily includes AI systems that generate or manipulate image, audio, or video content.
The proposed regulatory framework also establishes specific or stricter transparency obligations depending on the type of AI system.
These transparency obligations will apply particularly to AI systems that:
- Interact directly with individuals;
- Are used to recognise emotions or determine associations with social categories based on biometric data;
- Generate or manipulate content, including deepfakes.
ENFORCEMENT AND FINANCIAL PENALTIES
It is evident that the European Union intends to establish the necessary mechanisms to ensure the effective implementation and enforcement of this regulatory framework.
Failure to comply with the proposed rules may result in substantial financial penalties.
Depending on the nature of the infringement and the size of the company, the proposed penalties may reach EUR 35 million or 7% of a company’s total worldwide annual turnover, while other categories of infringements may be subject to lower percentage-based penalties, including 1.5% of worldwide annual turnover.
Undoubtedly, this regulatory framework represents an important step towards strengthening the European Union’s position in technology and innovation.
Initial steps towards regulating Artificial Intelligence have also been taken by China and the United States, although each jurisdiction has adopted a different approach to governing the legal relationships arising from the development and use of AI.
Google CEO Sundar Pichai has emphasised that Artificial Intelligence is “too important not to regulate well”, highlighting the significance of establishing an appropriate legal framework for this field.
Considering the rapid development of Artificial Intelligence and its present and future impact on people’s lives and business activities, the establishment of an effective and comprehensive regulatory framework remains of exceptional importance.
Author: Valmir Haziraj
