AI Act: European Union’s Artificial Intelligence Law Takes Effect

The European Artificial Intelligence Act (AI Act), the world’s first comprehensive regulation on artificial intelligence, came into effect on August 1, the European Commission has announced.

The AI Act aims to ensure that AI developed and used within the EU is trustworthy, with robust safeguards to protect fundamental rights. Additionally, the regulation seeks to create a unified internal market for AI across the EU.

“Today marks a significant milestone in Europe’s leadership in trustworthy AI,” said Thierry Breton, EU Commissioner for Internal Market.

The AI Act focuses on product safety and adopts a risk-based approach, categorizing AI systems into the following categories:

• Minimal Risk: Most AI systems pose minimal risk to citizens’ rights and safety. Companies may voluntarily adopt additional codes of conduct for these systems.
• Specific Transparency Risk: AI systems, such as chatbots, must disclose clearly to users that they are interacting with a machine. Additionally, users should be informed when biometric categorization or emotion recognition systems are in use.
• High Risk: AI systems classified as high-risk must adhere to strict requirements, including risk-mitigation measures. High-risk systems include those used for recruitment, determining loan eligibility, or operating autonomous robots.
• Unacceptable Risk: AI systems posing a clear threat to fundamental rights are banned. This category includes systems that manipulate human behavior to undermine free will.
• General-Purpose AI Models: The AI Act also introduces rules for general-purpose AI models, which are capable of performing a wide range of tasks, such as generating human-like text. The Act ensures transparency throughout the value chain and addresses potential systemic risks associated with these models.

Member States have until August 2, 2025, to designate national competent authorities responsible for overseeing the implementation of AI regulations and conducting market surveillance. However, the prohibition of AI systems deemed to present an unacceptable risk will come into effect six months earlier, starting in February 2025.

Companies that fail to comply with the regulations will face fines. Penalties may reach up to 7 percent of the global annual turnover for violations involving banned AI applications, up to 3 percent for breaches of other obligations, and up to 1.5 percent for providing incorrect information.