EU AI Act Takes Effect: A Landmark Legislation for Global Comprehensive AI Regulation

EU Artificial Intelligence Act Takes Effect: The World’s First Comprehensive AI Regulation

On August 1, 2023, the European Union officially enacted the Artificial Intelligence Act, marking the world’s first comprehensive regulatory framework for artificial intelligence. This landmark legislation aims to ensure the reliability of AI technologies developed and utilized within the EU while establishing a series of protective measures to safeguard fundamental human rights.

Thierry Breton, the EU Commissioner for the Internal Market, described the law as an “effective, balanced, and globally pioneering” framework, designed to create a unified AI market that fosters technological innovation and investment.

The Act introduces unprecedented penalties for non-compliance. Companies that violate prohibitions could face fines of up to 7% of their global annual revenue, while other infractions may incur penalties of up to 3%. Providing misleading information can lead to fines of up to 1.5%. These stringent measures are intended to deter violations and promote the healthy development of AI.

The legislation categorizes AI systems based on their risk levels, outlining specific regulatory requirements accordingly. For low-risk systems, such as recommendation algorithms and spam filters, which pose minimal threats to rights and safety, the Act imposes no specific obligations; however, companies are encouraged to voluntarily adopt additional best practices.

For systems with identifiable transparency risks, such as chatbots, the law mandates clear disclosure to users when they are interacting with machines, as well as labeling AI-generated content (e.g., deepfakes). Businesses employing biometric or emotion-recognition technologies must provide prior notifications to users. Furthermore, synthetic audio, video, and images must be labeled as machine-generated to enhance transparency and detectability.

High-risk AI systems, such as those used for recruitment assessments or autonomous robotics, are subject to stringent requirements, including the establishment of risk mitigation strategies, the use of high-quality datasets, maintaining activity logs, delivering detailed documentation and user information, and ensuring human oversight for robustness, accuracy, and cybersecurity.

The Act directly prohibits AI systems that pose clear threats to fundamental rights, including technologies that manipulate user behavior, toys that encourage minors to take risks, social scoring systems, and certain predictive policing applications. Strict limitations are also applied to emotion-recognition systems used in the workplace and real-time biometric systems utilized in law enforcement.

The implementation of the Artificial Intelligence Act represents a significant advancement for the EU in the realm of AI regulation, offering valuable insights and reference points for global AI governance. As the Act unfolds, the EU aims to create an AI ecosystem that encourages innovation while safeguarding rights.