Across its Generative AI white papers, the Financial Regulation Innovation Lab (FRIL) demonstrates how large language and multimodal models can be deployed as regulatory-aligned decision support tools across distinct but connected financial services challenges. In pensions and consumer support, FRIL examines how multimodal generative AI systems can deliver scalable targeted support while remaining within the FCA’s advice–guidance boundary, emphasising compliance-by-design, boundary discipline, auditability, and supervisory readiness through structured, explainable interactions (see Multimodal AI for Scaling Targeted Support :Navigating the FCA Advice–Guidance Boundary ). In ESG compliance, FRIL demonstrates how large language models, and their integration with large vision models, can simplify ESG reporting by extracting, synthesising, and generating analyst-style ESG narratives from unstructured corporate disclosures, while explicitly addressing both the capabilities and risks of Generative AI in regulatory settings (see Generative AI for Simplified ESG Reporting in Financial Services ). Extending these capabilities into regulatory foresight, FRIL shows how LLMs can be embedded into horizon scanning processes to detect early regulatory signals, analyse stakeholder feedback, and forecast regulatory evolution, using the FCA’s anti-greenwashing rule as a detailed case study (see LLM Horizon Scanning Anti-Greenwashing ). Taken together, these papers position Generative AI not as a standalone automation technology, but as a modular, auditable capability that supports compliance, supervision, and strategic preparedness across multiple financial services use cases.

Across its Explainable AI white papers, FRIL develops and applies explainability as a foundational capability for trustworthy, accountable, and regulatorily aligned AI deployment in financial services. In the context of financial risk management, FRIL sets out how Explainable AI can enhance the transparency, governance, and regulatory acceptability of AI-driven risk models, demonstrating its application to credit risk and default prediction while explicitly addressing the trade-off between model performance and explainability (see Explainable AI for Financial Risk Management). Building on this, FRIL introduces the concept of Explainable Intelligent Automation, positioning explainability as the critical interface between Robotic Process Automation, Business Process Management, and AI, and showing how transparent AI-enabled automation can simplify compliance processes while supporting organisational trust and adoption (see Simplifying Compliance through Explainable Intelligent Automation). A complementary set of papers focuses on fairness and algorithmic discrimination, developing a comprehensive, model-agnostic XAI toolbox grounded in ethical rights-to-explanation, and applying it to lending decisions to identify plausibly discriminatory outcomes and explore fairness–performance trade-offs under equality of opportunity and equality of outcome scenarios (see Promoting Fairness and Exploring Algorithmic Discrimination). This work is further extended to assess multiple protected characteristics simultaneously, demonstrating how XAI can support multidimensional fairness analysis while maintaining model performance and motivating human-in-the-loop interventions where warranted (see Fairness and Discrimination in Lending Decisions). Taken together, these papers position Explainable AI as a practical and necessary infrastructure for regulatory compliance, ethical assurance, and supervisory confidence in AI-enabled financial decision-making.

FRIL’s forthcoming work on Agentic AI, due in Q1 2026, will extend its existing research on multimodal generative AI and the advice–guidance boundary into a fully agentic setting. Building directly on the Multimodal AI for Scaling Targeted Support framework, the paper will develop and examine an agentic AI architecture centred on the Digital Pensions Assistant avatar demonstrator, where autonomous AI agents operate across multi-step interactions rather than single, static responses. The focus will be on how agentic systems plan, reason, and act over time within tightly defined regulatory constraints, and how autonomy, explainability, and accountability can be maintained in consumer-facing financial services use cases. By situating agentic AI within the FCA advice–guidance boundary, the paper will explore the implications of agent autonomy for compliance, supervision, and consumer protection, positioning Agentic AI as the next stage in FRIL’s research on scalable, regulator-aligned AI systems in financial services.

Across its Earth Intelligence white papers, FRIL develops a coherent Earth Intelligence framework for integrating Earth observation, geospatial data, and AI application to support regulatory compliance, ESG reporting, and supervisory oversight in financial services. The work begins by situating Earth Intelligence within the evolving EU sustainable finance architecture, examining how the EU Green Deal and the Sustainable Finance Framework create new data, disclosure, and verification demands that extend beyond traditional corporate reporting and into physically grounded environmental evidence (see The EU Green Deal and the Sustainable Finance Framework). Building on this regulatory foundation, FRIL analyses the European Sustainability Reporting Standards (ESRS) and identifies concrete opportunities for financial services to leverage Earth observation and geospatial datasets to evidence, validate, and enhance sustainability disclosures (see ESRS Opportunities for Financial Services). This is operationalised through a detailed mapping of ESRS disclosure datapoints to relevant Earth observation and geospatial datasets, demonstrating how satellite data, spatial indicators, and AI-enabled analytics can be systematically aligned with regulatory reporting requirements (see Mapping ESRS Disclosure Datapoints to Relevant Datasets). Extending Earth Intelligence beyond firm-level reporting, FRIL further demonstrates how geospatial data can be applied to supply chain assessment, enabling enhanced visibility of environmental and sustainability risks across complex, multi-tier value chains (see Supply Chain Assessment Using Geospatial Data). Taken together, these papers position Earth Intelligence as a critical evidential layer for sustainable finance, enabling regulators and financial institutions to move from narrative-based ESG claims toward spatially grounded, data-driven assurance.

Across its Open Finance and AI white papers, FRIL examines how data sharing frameworks and advanced analytics can support sustainability objectives, consumer resilience, and more effective financial support. One strand of this work focuses on how Open Finance-enabled data access can support the transition to carbon-neutral banking, emphasising the role of transaction-level financial data, consumer engagement, and measurement frameworks in assessing, monitoring, and influencing the carbon impact of banking activities (see Open Finance and Carbon Neutral Banking). Complementing this, FRIL’s paper From Crisis to Prosperity – AI and Open Finance, developed in collaboration with Sopra Steria, explicitly situates Open Finance and AI within the context of the FCA’s Advice Guidance Boundary Review (AGBR), framing “targeted support” as a regulatory mechanism to bridge the gap between generic guidance and full financial advice . The paper sets out how consent-driven Open Finance data, combined with explainable and fairness-aware analytical models, can enable persona-based, non-product-specific guidance that aligns with the AGBR and Consumer Duty frameworks, while supporting holistic financial health and resilience across spending, saving, borrowing, and long-term planning. In doing so, it positions Open Finance as the foundational data infrastructure that enables scalable, regulator-aligned targeted support, with AI acting as an enabling analytical layer rather than a substitute for regulatory judgement or human oversight. Taken together, these papers frame Open Finance as both a policy and data architecture through which more inclusive, resilient, and compliant financial support models can be delivered.

Across its work on Financial Crime Prevention and AI, the Financial Regulation Innovation Lab (FRIL) explores how data, analytics, and artificial intelligence can strengthen detection, mitigation, and information-sharing frameworks for fraud and financial crime, while remaining aligned with regulatory expectations and data governance principles. In collaboration with Sopra Steria, FRIL’s Enhancing Financial Crime Detection: AI Frameworks paper examines how machine-augmented detection models can support earlier and more accurate identification of suspicious activity, outlining a layered AI framework that integrates pattern recognition, anomaly detection, and explainability to augment human review and compliance workflows (see Enhancing Financial Crime Detection AI Frameworks). Complementing this, FRIL’s APP Fraud Mitigation: The Role of Data and Information Sharing paper focuses on the prevention of authorised push payment (APP) fraud through enhanced data sharing, cooperative intelligence, and joined-up information flows across firms and regulators, emphasising how consented, privacy-preserving data exchange can improve identification of fraud patterns without relying solely on advanced analytical models (see App Fraud Mitigation the role of Data and information sharing). Taken together, these papers position financial crime prevention as a domain where AI-boosted detection and cross-sector data sharing can operate in tandem to improve outcomes, while highlighting the importance of governance, explainability, and collaborative networks in maintaining trust and compliance across the financial system.

In its RegTech and AI white paper, FRIL examines how artificial intelligence can be integrated into regulatory technology to simplify compliance processes, enhance regulatory reporting, and improve supervisory effectiveness without compromising on governance or auditability. The paper outlines the key features of RegTech solutions - including automation, data standardisation, and real-time analytics - and situates AI as an enabling capability that can augment these features to reduce manual effort, improve accuracy, and support risk-based regulatory engagement (see Simplifying Compliance The Role of AI and RegTech) FRIL discusses how AI-enhanced RegTech can help firms meet increasingly complex reporting obligations, improve transparency for regulators, and facilitate more efficient supervisory interactions, while also emphasising the importance of explainability, data governance, and alignment with existing regulatory frameworks. The paper positions AI-enabled RegTech as a practical bridge between compliance obligations and operational efficiency in financial services.

In the area of ESG Greenwashing, FRIL examines how artificial intelligence can be used to enhance the measurement, detection, and mitigation of greenwashing in environmental, social, and governance (ESG) reporting, addressing both conceptual challenges and practical regulatory needs. The paper outlines how AI-enabled analytics — including natural language processing, pattern recognition, and multimodal data integration — can be applied to corporate disclosures, sustainability claims, and external environmental data to identify inconsistencies, unsupported assertions, or misleading representations that may constitute greenwashing (see ESG Greenwashing and Applications of AI for Measurement). It discusses the inherent difficulty of defining and operationalising “greenwashing,” the regulatory emphasis on accuracy and verifiability in ESG claims, and the potential for AI to provide scalable, data-driven evidence that complements human review and expert judgement. By positioning AI as a measurement and analytical tool - rather than a substitute for regulatory standards or human oversight - the paper highlights how AI can support more trustworthy sustainability reporting and more effective regulatory scrutiny in the context of rising scrutiny around ESG greenwashing.

In the area of Strategic Foresight, FRIL examines how formal scenario planning and strategic foresight methods can be harnessed by financial institutions and regulators to enhance future readiness in the face of technological, regulatory, and market uncertainty. The paper outlines the value of structured foresight - including horizon scanning, alternative futures, and stress-testing of strategic assumptions - as a means to anticipate disruptive trends, reduce cognitive biases, and support more resilient decision-making in fintech and broader financial services (see Strategic Foresight in Fintech Harnessing Scenario Planning for Future Readiness). Drawing on established foresight frameworks, it discusses how organisations can integrate scenario planning into strategic processes, align foresight outputs with risk management and innovation strategies, and use foresight to inform investment priorities, governance practices, and regulatory engagement. The paper positions strategic foresight not as a predictive tool, but as a disciplined practice that expands organisational awareness of plausible futures, supports adaptive strategy formulation, and strengthens capacity for anticipatory action in an increasingly complex financial ecosystem.

From Crisis to Prosperity - AI Open Finance

Supply Chain Assessment Using Geospatial Data

LLM Horizon Scanning Anti-Greenwashing

Strategic Foresight in Fintech Harnessing Scenario Planning for Future Readiness

App Fraud Mitigation the role of Data and information sharing

Simplifying Compliance The Role of AI and RegTech

Generative AI for Simplified ESG Reporting in Financial Services

Mapping ESRS Disclosure Datapoints to Relevant Datasets

The EU Green Deal and the Sustainable Finance Framework

ESRS Opportunities for Financial Services

Explainable AI for Financial Risk Management

Simplifying Compliance through Explainable Intelligent Automation

ESG Greenwashing and Applications of AI for Measurement

Promoting Fairness and Exploring Algorithmic Discrimination

Using Automation and AI to Combat Money Laundering

Enhancing Financial Crime Detection AI Frameworks

Fairness and Discrimination in Lending Decisions

Open Finance and Carbon Neutral Banking

In line with the demand-led principle on which FRIL is premised, the optimal impact from the white papers series will occur when there is close collaboration between academia and industry. We envisage a number of levels of engagement with industry on the white paper series. These are set out as follows, where the time commitments outlined are for guidance purposes:

Publication, Integrity and Ethics

FRIL will follow normal publishing conventions with white papers and with funded research (FRIL receives funding from Innovate UK, which is part of Research Councils UK).  This means we will publish under Creative Commons licensing, with papers having a digital object identification (DOI), and the overall FRIL papers series will have an ISSN number.  We will also follow norms of research integrity and research ethics.