Next month, the EU will introduce its groundbreaking AI legislation, the EU Artificial Intelligence Act, which aims to regulate AI to protect citizens from potential harm. While EU lawmakers are primarily concerned about consumer safety and the spread of deepfakes, the tech industry has criticized the legislation, calling it incomplete and stifling. The law classifies AI into different risk categories and imposes varying degrees of regulation, with minimal risk uses like video games exempted. High-risk applications, such as biometric identification and public service systems, will face stricter regulations. The legislation also bans AI systems that threaten citizens' rights, such as those used for deception or profiling.

The rules have faced challenges due to the emergence of generative AI models, and critics argue that the legislation lacks clarity, particularly regarding copyright and content responsibility. Compliance costs and the potential impact on small companies are also concerns. Tech companies have until February 2023 to comply with the "unacceptable risk" rules or face significant fines. Further secondary legislation is needed to outline implementation details, with a tight deadline.