The Financial Conduct Authority (FCA) has announced that Barclays, Experian, and UBS are among the latest financial institutions to join its live artificial intelligence (AI) testing initiative. The program, which began with a first cohort including Lloyds Banking Group, NatWest, and Monzo, provides a regulated environment where firms can experiment with AI applications in real-world conditions while receiving supervisory support. This second cohort marks a significant expansion of the initiative, reflecting the growing interest in deploying AI across the financial sector.
The FCA's sandbox-style approach allows participants to test customer-facing and business-to-business use cases. These include AI-powered investment advice, credit score insights for consumers, agentic payments, and anti-money laundering detection systems. According to the regulator, the firms are exploring a range of technologies from agentic AI and small language models to more experimental approaches such as neurosymbolic AI. The initiative is designed to help companies that have reached an advanced stage of AI development and are ready to implement their solutions in live markets, mitigating risks while fostering innovation.
Regulatory support and oversight
Jessica Rusu, the FCA’s chief data, information and intelligence officer, emphasized the collaborative nature of the project. “We’re continuing to collaborate with firms to support the safe and responsible development of AI in UK financial markets,” she said. She added that the “tailored” support provided by the FCA and its technical partner Advai demonstrates the regulator’s commitment to keeping pace with AI advancements and shows “how regulators and industry can work together to harness innovation responsibly.” The FCA stated that the initiative helps applicants explore key questions around risk management and live monitoring to ensure responsible deployment of AI for consumers and markets. The regulator plans to publish a report this year highlighting both good and poor practices observed during the trials.
The inclusion of major global banks like Barclays and UBS alongside credit bureau Experian underscores the breadth of AI use cases being considered. For instance, AI models can analyze vast amounts of transaction data to identify suspicious patterns more efficiently than traditional rule-based systems. Similarly, credit scoring algorithms powered by machine learning can provide more granular assessments of consumer creditworthiness, potentially expanding access to credit for underserved populations. However, these applications also raise concerns about bias, transparency, and data privacy, which the FCA’s regulatory oversight aims to address.
Background of AI regulation in UK financial services
The FCA’s AI testing initiative is part of a broader effort by UK regulators to understand and manage the implications of AI in finance. The Bank of England and the FCA have jointly established a framework for monitoring AI adoption, including the creation of a cross-regulatory AI working group. The UK government has also published a white paper on AI regulation, advocating a principles-based approach rather than a strict statutory regime. This has drawn criticism from some parliamentarians who argue that regulators have been too slow to act.
Earlier this year, the Treasury Committee of the House of Commons issued a report criticizing financial services regulators for what it described as a “wait-and-see” approach to AI regulation. The report stated that “the UK public and the country’s finance system are exposed to potential serious harm because regulators in the financial sector are not doing enough.” The committee’s chair, Meg Hillier MP, said that recent developments—such as Anthropic’s AI model Mythos, which uncovered decades-old vulnerabilities in banking systems—highlight the urgent need for proactive oversight. “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,” she said.
In response, Sarah Breeden, deputy governor for financial stability at the Bank of England, defended the central bank’s actions. “Far from taking a ‘wait-and-see’ approach, we have invested heavily in analysing the current and future risks posed by both the use of AI in financial services, and the broader investment in and adoption of AI across the wider economy,” she said. Breeden acknowledged that AI has “broad, complex and likely long-term implications for how the UK financial system serves the real economy,” but disagreed with the committee’s characterization of the bank’s stance.
The debate underscores a fundamental tension: regulators must balance the desire to foster innovation with the need to protect consumers and maintain financial stability. The FCA’s sandbox approach is one way to manage this tension, but critics argue that it is not enough. The Treasury Committee’s report recommended that regulators adopt more prescriptive rules, including mandatory testing of AI models before deployment and ongoing monitoring of their performance.
Historical context and comparisons
The UK has historically positioned itself as a global leader in fintech innovation, with the FCA’s regulatory sandbox—first launched in 2016—serving as a model for other countries. The AI testing initiative builds on this legacy by focusing specifically on the challenges posed by generative AI, agentic AI, and other advanced techniques. Similar initiatives exist in other jurisdictions, such as the U.S. Securities and Exchange Commission’s (SEC) emphasis on AI washing and the European Union’s AI Act, which imposes strict requirements on high-risk AI systems used in financial services.
The current second cohort includes firms that have made significant progress in AI development. Barclays, for example, has been investing heavily in machine learning for fraud detection and customer service automation. Experian leverages AI to improve credit scoring models, while UBS uses AI for wealth management advice and compliance monitoring. The FCA’s technical partner Advai provides expertise in adversarial testing and robustness evaluation, helping firms identify vulnerabilities in their AI systems before full-scale deployment.
One of the key use cases being tested in this cohort is agentic payments, where AI agents autonomously initiate and execute payments on behalf of customers. This technology could revolutionize how individuals and businesses manage cash flows, but it also introduces new risks around authorization, error correction, and fraud. Similarly, neurosymbolic AI—which combines neural networks with symbolic reasoning—promises to make AI decisions more interpretable, a critical requirement for regulated industries.
The FCA’s commitment to publishing a report on good and poor practices is intended to share lessons learned across the industry. Such transparency is vital for building trust in AI systems and ensuring that the benefits of innovation are realized without compromising consumer protection. As AI continues to evolve rapidly—exemplified by Anthropic’s Mythos model—the pressure on regulators to stay ahead of the curve will only intensify.
In the meantime, major UK banks are already in active discussions with regulators, finance ministries, and national security organizations following the discovery of vulnerabilities by the latest Anthropic model. These discussions highlight the interconnected nature of AI risks, where a model trained on public code repositories can uncover flaws in decades-old banking software. The Treasury Committee has called for immediate action, and the FCA’s sandbox provides a framework for testing mitigations before they are applied broadly.
Overall, the expansion of the FCA’s AI testing initiative signals that both industry and regulators recognize the transformative potential—and the inherent risks—of AI in finance. By bringing together diverse firms and use cases, the program aims to create a blueprint for responsible AI adoption that could serve as a benchmark for the global financial sector.
Source: ComputerWeekly.com News