The Bank of England has called for the development of a dedicated regulatory framework for artificial intelligence, arguing that existing technology oversight mechanisms are insufficient to address the unique risks posed by increasingly autonomous AI systems in financial services.
In remarks delivered this week, a senior BoE official outlined the central bank’s position that agentic AI—systems capable of autonomous decision-making and action—requires regulatory approaches distinct from traditional software or algorithmic trading rules. The intervention represents a notable departure from the UK government’s current light-touch, sector-agnostic approach to AI governance.
The timing reflects mounting concern among financial regulators about AI systems that can initiate transactions, adjust risk parameters, or interact with customers without human oversight. Unlike conventional automated systems that follow predetermined rules, agentic AI can adapt its behaviour based on environmental feedback, creating novel supervisory challenges for institutions accustomed to auditing static code and fixed decision trees.
The BoE’s position contrasts with the European Union’s AI Act, which takes a horizontal approach across all sectors, and with the United States’ fragmented regulatory landscape where AI oversight remains distributed among existing financial regulators. The central bank appears to be advocating for a middle path: sector-specific rules that account for AI’s distinctive characteristics whilst remaining tailored to financial services’ systemic importance.
This stance carries significant implications for the UK’s financial services sector, which employs AI across credit decisioning, fraud detection, trading operations, and customer service. Major banks including HSBC, Barclays, and NatWest have invested heavily in machine learning capabilities, with the sector accounting for an estimated 15-20 per cent of enterprise AI spending in the UK economy.
Financial institutions face a delicate balancing act. Bespoke regulation could provide clearer compliance pathways and reduce legal uncertainty compared to applying general data protection or consumer duty rules to AI systems. However, new requirements would likely mandate additional testing, documentation, and human oversight mechanisms, potentially slowing deployment timelines and increasing operational costs.
Technology vendors serving the financial sector—including established players such as IBM and SAS alongside AI-native firms like Dataiku and H2O.ai—would need to adapt their products to meet sector-specific standards. This could raise barriers to entry for smaller firms whilst creating opportunities for specialised compliance and auditing services.
The BoE’s intervention also positions London as a potential standard-setter for AI governance in finance. With the EU’s comprehensive framework already in force and US regulators moving slowly, a well-designed UK approach could attract firms seeking regulatory clarity without Brussels’ prescriptive requirements. Conversely, divergent standards across jurisdictions could fragment the market for AI services and complicate operations for global financial institutions.
The central bank has not yet published detailed proposals, leaving critical questions unanswered. These include whether regulation would focus on AI development practices, deployment contexts, or specific use cases; how requirements might scale based on system autonomy or potential impact; and whether oversight would sit with the BoE, the Financial Conduct Authority, or a new body.
The call for bespoke regulation follows similar moves by the Prudential Regulation Authority, which has been examining AI risks in banking supervision, and the FCA’s work on algorithmic trading oversight. It also aligns with recommendations from the Treasury’s AI in Financial Services review, though the government has yet to commit to legislative action.
Industry response will likely prove mixed. Large institutions with established compliance functions may welcome clear rules, whilst smaller banks and fintech firms could view additional requirements as competitive disadvantages. Trade bodies including UK Finance and the CityUK will probably push for proportionate approaches that preserve innovation capacity.
The immediate focus will be whether the BoE publishes a consultation paper outlining specific regulatory proposals and whether the Treasury incorporates AI-specific measures into upcoming financial services legislation. How the central bank defines ‘agentic’ AI and sets thresholds for enhanced oversight will determine whether this represents incremental adjustment or fundamental reorientation of technology governance in finance.
