EU AI Act Update: First Provisions Are Live
The first provisions of the EU AI Act came into effect on 2 February 2025.
Prohibited AI Practices
Central to the EU AI Act is Article 5, which outlines a list of prohibited AI practices. These include:
The use of facial recognition databases created from images scraped online or captured via security footage.
Biometric systems for identity verification.
Manipulative techniques designed to influence human behaviour.
Social scoring systems.
Criminal prediction software.
Emotional detection technologies in educational and workplace settings.
AI systems that exploit an individual's age, disability, or socioeconomic status to manipulate behaviour.
Guidance and AI Literacy
Despite the absence of immediate guidance from the European Commission, stakeholders are expected to adapt to these new regulations. The Commission plans to release nonbinding guidelines based on a consultation launched in November 2024, with the aim of providing clarity on the application of these prohibitions "in early 2025." Additionally, a repository of AI literacy practices will be compiled to ensure that those deploying AI systems possess the necessary technical knowledge and understanding. For now, check RMOK Legal’s article: AI Literacy - What’s your plan? Need a suggestion?!?
Emphasis on AI Literacy
Article 4 of the act emphasises the importance of AI literacy, requiring users within companies to be well-versed in the functionalities and implications of the AI systems they employ. This focus on literacy acknowledges the profound impact AI can have on society. The European Commission will host a webinar on 20 February to further explain these requirements, with the EU AI Office playing a key role in their implementation.
Future Provisions and Enforcement
The most consequential aspects of the AI Act are set to take effect in 6 months. On 2 August, member state authorities will be empowered to enforce the regulation and issue fines, marking the beginning of a new era of accountability in AI governance. The list of prohibited practices will undergo an annual review to ensure the act remains responsive to the evolving landscape of AI technology. Additionally, obligations for providers of general-purpose AI models will come into force, further solidifying the EU's commitment to responsible AI development.
In case you missed it, check out our insights:
#EUAIAct #AIregulation #AIethics #ArtificialIntelligence #AISafety #Compliance #TechLaw #DataGovernance #RiskManagement #DigitalSingleMarket #Innovation #Startups #Europe #AIcompliance #EthicalAI #AIforGood #FutureofAI #RMOKLegaI