Europe and the United Kingdom entered the summer of 2025 amid significant developments in Artificial Intelligence governance, regulatory planning, and data law reform. Anticipation rose across industry as the EU finalized the timeline for enforcing sector-shaping rules under the EU Artificial Intelligence Act, especially for providers of general purpose Artificial Intelligence models (GPAI). With compliance obligations to take effect from 2 August 2025 and enforcement—with potential fines—to follow by August 2026, regulatory scrutiny on global Artificial Intelligence providers is increasing. The recent publication of the GPAI Code of Practice and detailed guidelines from the European Commission seeks to clarify industry obligations after much delay and uncertainty. The Code divides compliance expectations into key areas around transparency, copyright, and safety, formalizing commitments and documentation frameworks for GPAI providers, with an eye towards reducing administrative burden and clarifying legal standing. Notably, earlier-placed models benefit from transitional timelines, reflecting a phased regulatory approach.
The EU’s wider Artificial Intelligence strategy is laid out in the AI Continent Action Plan, an ambitious roadmap stressing compute infrastructure expansion—including plans for AI Gigafactories and major GPU acquisition—alongside enhanced data sharing (the Data Union Strategy) and industry-focused Artificial Intelligence adoption initiatives. Regulatory simplification, workforce upskilling, and structured industry consultations all feature prominently, signaling Europe´s bid to lead both technically and ethically. Complementing legislative action, new standards bodies are shaping practice: the European Telecommunications Standards Institute launched its TS 104 223 security specification, positioned as the first global standard for securing Artificial Intelligence across its lifecycle. The European Data Protection Board released a major report on privacy in Large Language Models, emphasizing deep risk assessment, data lifecycle mapping, and early-stage governance. Meanwhile, the European Union Intellectual Property Office published a 400-page analysis on generative Artificial Intelligence and copyright, exposing fragmentation in opt-out mechanisms and underscoring the urgent need for common, enforceable frameworks.
In the United Kingdom, 2025 was marked by legislative restraint and strategic delay. The landmark Data (Use and Access) Act updated UK data privacy and addressed smart data schemes, but sidestepped direct Artificial Intelligence copyright provisions—after intense parliamentary debate—pending further consultation with the Intellectual Property Office. Instead, a new economic impact analysis and technical report on copyright and Artificial Intelligence must be delivered within nine months of Royal Assent, with progress required by year-end. The UK government has postponed its headline Artificial Intelligence Bill until summer 2026, citing the desire for transatlantic alignment and bringing copyright concerns under one umbrella. The Information Commissioner’s Office published its first comprehensive Artificial Intelligence and Biometrics Strategy, targeting high-risk use cases—such as foundation model development and automated decision-making—and promising a future statutory code of practice. Cross-sector collaborations, digital policy toolkits, and a renewed focus on regulatory clarity reflect ongoing efforts to encourage innovation without sacrificing trust or data protection. As both the EU and UK seek to balance innovation incentives against pressing regulatory, security, and rights-protection challenges, the competitive landscape for Artificial Intelligence governance is rapidly evolving, with key consultations and further guidance expected before mandatory compliance arrives in 2026.
