The EU AI Regulation: Phased implementation and upcoming obligations
Ioana Hategan
by Ioana Hategan
The European Union's landmark AI Regulation is advancing its phased implementation, harmonising rules for artificial intelligence (AI) while safeguarding fundamental rights and fostering innovation. Designed to improve the internal market's functionality, the Regulation promotes human-centric and trustworthy AI while addressing the risks posed by emerging technologies.
The Regulation's phased approach ensures readiness among stakeholders, focusing first on provisions related to AI literacy, prohibitions, and governance. Chapters I and II take effect on 02 February 2025, emphasising the need for AI literacy and banning certain harmful AI practices. These practices include social scoring, manipulative AI, real-time biometric identification in public spaces (with limited exceptions), facial recognition scraping, and emotion recognition in sensitive contexts such as workplaces and educational institutions. Such bans reflect the EU's commitment to upholding European values, privacy, and dignity.
Non-compliance with these prohibitions will result in severe penalties, including fines of up to EUR 35 million or 7% of global turnover, effective 02 August 2025. Additionally, EU member states must establish governance structures and enforcement mechanisms by this date, ensuring that penalties are applied consistently across the EU.
Key governance measures, effective from August 2025, require each member state to designate notifying authorities responsible for assessing and monitoring conformity assessment bodies (CABs). These CABs ensure AI systems meet the Regulation’s stringent safety and quality requirements. Governance structures, including the AI Office and European AI Board, will facilitate collaboration among member states, harmonise enforcement, and provide updated guidance in response to technological advancements.
Provisions for general-purpose AI models (GPAI), particularly those posing systemic risks, will also be enforced from August 2025. Providers of GPAI must meet heightened obligations, such as conducting risk assessments, maintaining transparency, and addressing potential societal impacts. These measures aim to minimise misuse while fostering innovation, ensuring that GPAI systems are developed responsibly.
The Regulation's full effect will be realised on 02 August 2026, when most provisions will become enforceable. However, obligations for operators of high-risk AI systems under Article 6(1) will be delayed until 02 August 2027, providing additional time for adaptation.
By addressing risks, promoting transparency, and fostering innovation, the EU AI Regulation positions the EU as a global leader in trustworthy AI. As implementation progresses, stakeholders must act proactively to comply with upcoming obligations, ensuring the Regulation's goals of a safer and more innovative AI ecosystem are achieved.
For tailored legal advice on the Regulation and its implications, please contact our office. Our team is equipped to guide you through this complex framework, ensuring compliance and success in the evolving AI landscape.
Hategan Attorneys is an independent, full-service business law firm located in Timisoara, Romania, providing a tailored and integrated range of legal services to meet the needs of a wide variety of businesses, individuals, and organisations. More than three quarters of their clients are international companies which plan to invest, or are already doing business in Romania.
GGI member firmHategan AttorneysTimisoara, RomaniaT: +40 256 430 454
Law Firm Services
As Founding Partner of Hategan Attorneys, Ioana Hategan is responsible for managing the firm and is the Leader of the M&A and Private Equity Practice. Contact Ioana.
