EU Delays High-Risk AI Act Rules in Major Regulatory Shift

The European Commission has unveiled a major “Digital Omnibus” package, streamlining various digital regulations and notably postponing several high-risk provisions of the AI Act until December 2027.

These delayed provisions include AI applications in sensitive areas such as credit scoring, loan evaluations, and biometric identification.

The move is framed by Brussels as a pragmatic reprieve: regulators argue that implementation support tools (like technical standards and audit frameworks) still need to mature to ensure compliance is viable.

Why It Matters: Financial institutions, particularly lenders and credit providers, are being given extra time to embed compliance into their AI systems. But some risk still remains: more runway for innovation, yes, but also the potential for weaker consumer protections if the high-risk designation is softened.

GaiaLens Take: This isn’t deregulation so much as a recalibration. Firms should use this breathing space strategically: don’t delay your AI governance build-out just because the deadlines are loosening.

France & Germany Back Simplification of EU Digital & AI Rules

France and Germany have publicly backed the European Commission’s simplification plan, sending a strong signal in support of the Digital Omnibus. Under the proposed changes, some of the AI Act’s “high-risk” rules may move from rigorous third-party assessments to self-assessments, potentially reducing compliance burdens.

Supporters argue the simplification will free up capital and reduce red tape, particularly for businesses deploying AI in sectors like finance.

Why It Matters: If adopted, these reforms could drastically lower the compliance cost for financial institutions working with AI, especially in use cases like credit underwriting or risk modelling. But it also increases the burden on firms to self-police effectively.

GaiaLens Take: This is a bet on trust: the EU seems to be leaning toward a lighter, more flexible regime, but firms must not mistake flexibility for laxity. Robust internal governance will become the first line of regulatory defence.

UK Banks Struggle to Scale AI; Legacy Systems Dragging Behind

New research from CFOtech reports a clear mismatch in the UK banking industry: 80% of senior banking leaders believe their institutions are well-positioned for AI, but only 55% have actually deployed it in meaningful ways.

The biggest obstacles seem to be outdated core banking systems, poor data infrastructure, and regulatory uncertainty, according to the survey.

Many bankers compare running AI on legacy systems to "fuelling an EV with petrol," suggesting that without core modernisation, AI’s full promise simply can’t be realised.

Why It Matters: The gap between ambition and execution could become a bottleneck for UK banks. Without upgrading the plumbing, an AI-driven transformation may remain superficial.

GaiaLens Take: To lead in the AI era, banks must do the hard work first: modernise data foundations and core systems. Otherwise, they risk being outpaced by more agile competitors who built cloud-native from the ground up.

Bank of England Deploys Generative AI to Power Internal Supervision

The Bank of England is moving beyond theory; it is now using generative AI internally for tasks such as meeting summaries, document review, and compiling reports.

More ambitiously, the BoE is developing AI tools for its supervisory work, leveraging cloud-based large language models (LLMs) to analyse large volumes of unstructured data from regulated financial firms.

These deployments are guided by the BoE’s own internal governance framework, including a committee dedicated to AI ethics, risk testing, and durability.

Why It Matters: The BoE is doing more than regulating; it’s learning by doing. Its hands-on experience with generative AI could give it a unique vantage point on systemic risk, model failures, and governance challenges.

GaiaLens Take: This is a powerful signal. When a central bank uses AI internally, it’s not just about efficiency; it’s about risk literacy. Financial institutions should watch closely: BoE’s experience may well set next-gen regulatory expectations.

UBS Hires Chief AI Officer as It Doubles Down on Generative & Agentic AI

UBS has announced that Daniele Magazzeni will join as its Chief Artificial Intelligence Officer from 1st January 2026. Magazzeni joins from JPMorgan (where he led analytics for EMEA) and brings a mix of institutional AI deployment experience and academic research credentials.

He will lead the newly formed Chief AI Office and oversee UBS’ “Big Rocks” AI transformation program, which already includes over 300 active AI use cases, including generative tools and agentic agents.

Why It Matters: UBS’s commitment to a C-suite AI role reflects how seriously major banks are taking advanced AI. It’s not just a back-office tool; it’s becoming a strategic lever across operations, front-office, and client experience.

GaiaLens Take: This hire is more than symbolic. UBS is clearly betting that generative and agentic AI will drive its next wave of transformation, and by centralising AI leadership, it’s putting governance, value creation, and scale on equal footing.

Final Thought

This week’s headlines reflect a clear turning point: AI in finance is maturing from experimentation to structure. Whether through regulatory flexibility, internal deployment, or executive leadership, financial institutions are laying the foundations for long-term, scalable AI. The firms that build trust and capability now will define what finance looks like in the next decade.