EU AI Act First Enforcement Actions Target Prohibited AI Practices

On February 2, 2025, the European Union's Artificial Intelligence Act entered its first enforcement phase, marking a watershed moment in global AI regulation. The initial provisions that took effect included comprehensive prohibitions on AI systems deemed to pose unacceptable risks to fundamental rights and safety. Specifically banned were AI systems for social scoring by public authorities, real-time biometric identification in public spaces (with limited exceptions), emotion recognition in workplaces and educational institutions, biometric categorization systems inferring sensitive personal data, and AI systems exploiting vulnerabilities of specific groups. Within weeks of the enforcement date, the European Commission and member state authorities began coordinating investigations into suspected violations. Companies across multiple sectors found themselves under scrutiny for existing AI deployments that suddenly fell under the prohibited categories. The enforcement focused initially on the most egregious cases, including workplace surveillance systems using emotion recognition technology and retail environments deploying unauthorized biometric categorization. Several multinational technology companies received formal investigation notices regarding their AI systems' compliance status. The compliance challenges proved substantial for organizations that had not adequately prepared for the AI Act's requirements. Many companies discovered that AI systems they considered routine business tools actually fell under the Act's prohibited categories. The complexity of determining which specific AI applications were banned created significant interpretation challenges, particularly around the boundaries of emotion recognition and biometric categorization. Legal and compliance teams struggled to audit existing AI deployments and determine which systems required immediate discontinuation. The enforcement actions had immediate global implications, as multinational companies operating in the EU were forced to reassess their AI practices worldwide. Several major technology platforms announced they would disable certain AI features in EU markets rather than risk violations. The ripple effects extended beyond Europe, with companies in other jurisdictions proactively reviewing their AI systems to avoid similar regulatory challenges. Industry associations called for clearer guidance on compliance requirements, while some companies sought safe harbor provisions through voluntary compliance programs. The first phase of EU AI Act enforcement established a new baseline for acceptable AI practices that influenced regulatory discussions in the United States, United Kingdom, and other major markets.