This Sunday, February 2, key provisions of the EU AI Act will take effect, including the ban on AI systems deemed to pose an "unacceptable risk" and the deadline for meeting AI literacy requirements.

From this date, the use or marketing of such prohibited AI systems will be illegal within the EU, with penalties reaching up to €35 million or 7% of global annual turnover for non-compliance. Prohibited activities include manipulative or deceptive AI techniques, social scoring, certain biometric categorization practices, and real-time remote biometric identification in public spaces for law enforcement, among others. The Act’s broad scope means that even non-EU companies with AI operations affecting EU users will be subject to these regulations.

Beyond prohibitions, the Act also imposes literacy requirements, ensuring organizations provide adequate training and upskilling for teams handling AI systems. Commentators highlight the global impact of these rules, especially as other jurisdictions, like the US, take a more hands-off regulatory approach. Despite this divergence, US-based firms operating in the EU must still comply. Legal experts stress that businesses should use this deadline as a starting point for broader compliance efforts, particularly as the rest of the Act will roll out over the next 18 months. With many organizations lacking centralized records of AI deployments, identifying and assessing AI systems will be a complex but necessary step in preparing for full compliance.