AI Regulation and the Regulated Enterprise — Trajectory to 2030

Narrative

The regulatory landscape for Artificial Intelligence in Europe is rapidly evolving, with the EU AI Act establishing a comprehensive, risk-based framework. Key obligations for General-Purpose AI (GPAI) models, including transparency and copyright rules, became applicable in August 2025, with the European Commission's enforcement powers commencing in August 2026. Academic research highlights the technical challenges for generative AI providers to meet transparency obligations under Article 50 II by August 2026, suggesting that compliance requires fundamental architectural design rather than superficial labelling.

Further academic analysis explores the allocation of responsibility and liability across the complex GPAI value chain, identifying potential gaps in existing EU instruments. While the AI Act aims for a human-centric approach, embedding fundamental rights as legal thresholds, some scholars point to accountability deficits for autonomous AI agents, particularly in critical infrastructure, where certain provisions may be narrowed. The 'AI omnibus' simplification package, politically agreed in May 2026, aims to streamline implementation, extending transition periods for high-risk AI systems to 2027 and 2028, signalling an effort to balance regulatory ambition with practical feasibility.

In contrast, the UK maintains a pro-innovation, contextual approach, relying on existing sectoral regulators like the FCA, PRA, and Bank of England to interpret cross-sector principles. A 2023 feedback statement from these authorities confirmed a preference for technology-neutral, outcomes-based regulation over a specific AI definition. However, a January 2026 Treasury Committee report criticised this 'wait-and-see' stance, urging regulators to provide clearer guidance on applying existing rules, conduct AI-specific stress testing, and designate major AI and cloud providers as critical third parties by the end of 2026.

The FCA launched its Mills Review in January 2026, examining the long-term impact of AI, including autonomous and agentic systems, on retail financial services and market structure. Despite political pressure, UK financial regulators reiterated their commitment to a technology-agnostic approach in April 2026, while acknowledging emerging risks from general-purpose AI tools at the edge of their regulatory perimeter. Meanwhile, research from METR, utilising benchmarks like HCAST and RE-Bench, continues to empirically evaluate the autonomous capabilities and R&D potential of frontier AI models, providing quantitative insights into their rapidly increasing ability to complete complex tasks.