The Financial Conduct Authority (FCA) has expanded its artificial intelligence (AI) testing initiative, welcoming Barclays, Experian and UBS into the second cohort of participants. These firms will join earlier entrants such as Lloyds Banking Group, NatWest and Monzo in a supervised, real-world environment designed to foster responsible AI innovation within UK financial markets.
Expanding the AI Sandbox
The FCA's AI testing initiative, first launched in 2024, offers a controlled space where financial firms can deploy AI applications under regulatory oversight. This sandbox approach allows companies to test advanced technologies — including agentic AI, small language models (SLMs) and neurosymbolic AI — in live conditions while receiving tailored support from the regulator and its technical partner, Advai. The second cohort focuses on customer-facing and business-to-business use cases, such as AI-enabled investment support, consumer credit score insights, agentic payment systems and money laundering detection.
Jessica Rusu, chief data, information and intelligence officer at the FCA, stated: "We’re continuing to collaborate with firms to support the safe and responsible development of AI in UK financial markets." She emphasised that the tailored support reflects the regulator's commitment to keeping pace with AI advancements and demonstrates how regulators and industry can work together to harness innovation responsibly. The FCA added that the initiative helps applicants explore key questions around risk management and live monitoring, ultimately supporting the responsible deployment of AI for consumers and markets.
Background and First Cohort
The first cohort, announced in early 2025, included major high street banks and digital challengers. Lloyds Banking Group, NatWest and Monzo were among the initial participants, testing AI applications ranging from fraud detection to personalised financial advice. The programme was designed to bridge the gap between theoretical AI development and practical, compliant market deployment. The FCA noted that banks are increasingly experimenting with cutting-edge technologies: agentic AI systems that can autonomously execute tasks, small language models that offer efficiency and lower computational costs, and neurosymbolic AI that combines neural networks with symbolic reasoning for more explainable outcomes.
Regulatory Scrutiny and Criticism
Despite the FCA's proactive sandbox, UK financial regulators have faced criticism from parliamentary bodies. Earlier this year, the Treasury Committee accused the FCA and the Bank of England of adopting a "wait-and-see" approach to AI regulation. In a report, MPs warned that the UK public and the country's financial system are exposed to potential serious harm because regulators are not doing enough. The committee chair, Meg Hillier MP, stated: "Recent developments in the world of AI, such as Anthropic’s Project Mythos, show us how fast this transformative technology is moving. It has never been more important that those responsible for maintaining the UK’s financial stability take a proactive approach to understanding and mitigating the risks AI may pose to our financial system."
In response, Sarah Breeden, deputy governor for financial stability at the Bank of England, pushed back against the characterisation. She said: "We share the committee’s view that AI has broad, complex and likely long-term implications for how the UK financial system serves the real economy. However, we do not agree with its characterisation that the bank is taking a ‘wait-and-see’ approach." Breeden highlighted that the Bank has invested heavily in analysing current and future risks posed by AI in financial services, as well as the broader adoption of AI across the economy. She added that the Bank's work includes stress testing, horizon scanning and collaboration with international regulators.
The Anthropic Mythos Incident
The recent controversy surrounding Anthropic's AI model, Mythos, has intensified calls for stronger oversight. Last week, major UK banks entered discussions with regulators, finance bodies and national security organisations after Mythos unearthed decades-old software vulnerabilities. The incident demonstrated how advanced AI can identify weaknesses that human auditors might miss, but also raised concerns about the potential for such models to be used maliciously or to cause unintended disruptions. The FCA has acknowledged that the speed of AI evolution requires constant vigilance and adaptive regulatory frameworks.
Use Cases and Technology
The second cohort of the AI testing initiative will explore a variety of real-world applications. Barclays is expected to test agentic payments — AI systems that can initiate and manage transactions autonomously based on predefined rules and real-time data. Experian will focus on providing consumers with deeper credit score insights, possibly using natural language processing to explain credit decisions. UBS is likely to experiment with AI-driven money laundering detection, leveraging pattern recognition and anomaly detection to identify suspicious transactions more effectively. These use cases reflect the broader trend of financial institutions adopting AI to improve operational efficiency, enhance customer experience and strengthen compliance.
From a technological perspective, the FCA noted that participants are exploring a spectrum of AI models. Agentic AI represents a shift from passive recommendation systems to autonomous agents capable of executing complex workflows. Small language models offer a more resource-efficient alternative to large language models (LLMs), making them suitable for on-device processing and real-time applications. Neurosymbolic AI, an emerging field, aims to combine the learning capabilities of neural networks with the reasoning power of symbolic AI, potentially offering greater transparency and explainability — a critical requirement in regulated financial environments.
Global Context and Best Practices
The FCA's initiative is part of a broader global trend. Regulators in the European Union, Singapore and the United States are also developing sandboxes and frameworks to manage AI in finance. The EU's AI Act, which classifies AI applications by risk level, will have implications for financial firms operating in Europe. The FCA's approach — offering a safe space for experimentation while publishing guidance on good and poor practices — aims to position the UK as a leader in responsible AI innovation. The regulator has committed to publishing a report later this year highlighting best practices and lessons learned from both cohorts. This report is expected to inform future regulatory policy and help firms navigate the complex intersection of AI and financial compliance.
Risk Management and Monitoring
A key focus of the sandbox is risk management. Participants must demonstrate robust monitoring systems to detect and mitigate AI-related risks, such as bias, model drift, data privacy breaches and operational failures. The FCA's technical partner Advai provides expertise in AI assurance, helping firms validate their models against regulatory standards. The sandbox also facilitates knowledge sharing among participants, allowing firms to learn from each other's successes and challenges. This collaborative approach is intended to accelerate the safe deployment of AI across the financial sector while reducing the likelihood of systemic failures.
Critics, however, argue that sandboxes may not be sufficient to address the broader risks posed by AI. The Treasury Committee's report called for more stringent regulations, including mandatory stress testing and incident reporting. The Bank of England has responded by developing a framework for AI stress testing that will simulate scenarios such as market disruptions caused by AI trading algorithms or cyberattacks enabled by AI. These measures, combined with the FCA's sandbox, represent a multi-layered approach to AI governance.
Looking ahead, the FCA plans to expand the initiative to include more firms and a wider range of use cases. The regulator is also exploring ways to integrate AI oversight into its broader supervisory activities, such as the use of AI tools to monitor market abuse and consumer protection. As AI continues to evolve, the balance between fostering innovation and ensuring safety will remain a central challenge for financial regulators worldwide. The FCA's sandbox, while not a complete solution, provides a valuable testing ground for the technologies that will shape the future of finance.
Source: ComputerWeekly.com News