EU Artificial Intelligence Law Enters into Force

With the Artificial Intelligence Act (AI Act), the European Union (EU) aims to establish a comprehensive legal framework for the development, marketing and use of artificial intelligence in accordance with EU values. This framework aimed to promote the adoption of human-centered and trustworthy AI, while ensuring a high level of protection of health, safety and fundamental rights, including democracy, the rule of law and environmental protection. For these purposes, further to inter-institutional negotiations, on 13 March 2024, the European Parliament approved the AI Act; taking one of the significant steps that will determine the future of artificial intelligence in the world.

At this point, on 12 July 2024, the AI Act was published in the EU Official Journal. The law will enter into force on 1 August 2024.

The AI Act defines artificial intelligence as a machine-based system designed to operate at various levels of autonomy. AI Act will apply to (i) providers of artificial intelligence systems to be offered on the EU market, regardless of whether they are based in the EU, (ii) users of artificial intelligence systems located in the EU, and (iii) providers and users of non-EU artificial intelligence systems if they are used in the EU.

We have compiled the significant regulations introduced by the AI Act below:

Respect for sensitive data and privacy protection: AI Act introduces regulations to ensure that artificial intelligence systems to be used in EU countries are secure and respect fundamental rights.

Risk-based approach and regulations: AI Act introduces rules based on the risk level of artificial intelligence systems. In other words, the higher the risk, the stricter the rules to be applied. In this context, examples of high-risk AI use cases include critical infrastructure, education and vocational training, employment, basic private and public services, certain systems in law enforcement, and rules that closely concern public order such as immigration.

Regulations to encourage and support innovation: AI Act introduces national- level regulatory provisions for SMEs and entrepreneurs. The law will enable the development and testing of innovative artificial intelligence products within an ethical framework.

Prohibited practices and ethical principles: AI Act envisages several prohibited practices that aim to promote the ethical use of AI-based applications. Some of these practices are as follows:

— Analyzing the emotions of those concerned in business and educational settings,

— Crime prediction based on character analysis,

— Use of applications aimed at manipulating human behavior,

— Biometric categorization systems using sensitive characteristics (e.g. religious, philosophical beliefs, sexual orientation, race), and

— Untargeted scraping of facial images from the internet or CCTV images to create facial recognition databases.

In case of non-compliance with the rules introduced by the AI Act, the penalties to be paid by companies will vary depending on the size of the violation and the company.

The European Union Artificial Intelligence Office is a new unit established by the European Commission to coordinate the implementation, execution and harmonization activities of the AI Act. The AI Office will supervise AI systems based on general purpose models and systems provided by the same provider and will have the powers of a market surveillance authority. The office will coordinate governance among member states and oversee the implementation of rules on general-purpose artificial intelligence. Member state authorities will determine enforcement rules, including fines and other enforcement measures, warnings and non-monetary measures. Individuals will be able to file infringement complaints with the national competent authority, which will be able to initiate market surveillance activities.

In case of non-compliance with the rules introduced by the AI Act, it is envisaged that the fines to be paid by companies will vary between 35 million Euros or 7% of global turnover and 7.5 million Euros or 1.5% of turnover, depending on the violation and the size of the company.

The AI Act will gradually become fully enforceable within 24 months from 1 August 2024, the date of entry into force, as follows:

— On 2 August 2024, EU Member States will phase out prohibited systems;

— On 2 August 2025, obligations on multi-purpose artificial intelligence systems will enter into force;

— On 2 August 2026, the obligations for high-risk systems defined in the list of high-risk use cases will enter into force; and

— On 2 August 2027, high-risk system obligations that are currently subject to other EU legislation will enter into force.

This is one of the most important steps taken for AI-based technologies. In addition to being a concrete step, the AI Act also provides a comprehensive roadmap to ensure that AI technologies are developed and used safely and ethically. Therefore, it is critical for companies operating in the artificial intelligence sector to initiate the process of harmonizing with the systematics envisaged by the AI Act.

The full text of the AI Act is available here (available in EU languages).