Large-scale language models (LLMs) have the potential to improve efficiency and safety in the financial sector by detecting fraud, generating financial insights, and automating customer service, according to research from the Alan Turing Institute.
LLM's ability to rapidly analyze large amounts of data and generate coherent text is increasing understanding of its potential to improve services in a variety of sectors, including healthcare, law, education, and financial services such as banking, insurance, and financial planning.
The report, the first to explore the adoption of LLMs across the financial ecosystem, shows that those working in the sector have already started using them to support a range of internal processes, such as regulatory review, and are evaluating their potential to support external activities such as providing advisory and trading services.
In parallel with the literature review, the researchers organized a workshop that brought together 43 experts from major leading banks, investment bankers, regulators, insurance companies, payment service providers, government and legal experts.
The majority of workshop participants (52%) are already using these models to improve performance on information-oriented tasks, from meeting note management to cybersecurity and compliance insights, while 29% are using them to improve their critical thinking skills and a further 16% to break down complex tasks.
Systems are already being built in this field to rapidly analyze large amounts of text to streamline decision-making processes and risk profiling, improving investment research and back-office operations and thereby increasing productivity.
When asked about the future of the LLM in finance, participants felt that within two years, the LLM would be integrated into services such as investment banking and venture capital strategy development.
It is also likely that LLMs will be integrated to improve human-machine interaction, for example through dictation and built-in AI assistants to reduce the complexity of knowledge-intensive tasks such as regulatory review.
However, participants also acknowledged that the technology carries risks that limit its use: Financial institutions are subject to extensive regulatory standards and obligations that limit their ability to use AI systems that cannot produce outputs that are unexplainable, predictable, consistent, and without risk of error.
Based on their findings, the authors recommend that financial services professionals, regulators, and policymakers collaborate across the sector to share and develop knowledge on the implementation and use of LLMs, particularly as it relates to safety concerns. They also suggest that the growing interest in open source models should be explored, and that while they can be used and maintained effectively, mitigating security and privacy concerns should be a top priority.
Lead author Professor Kirsten Maple, Turing Fellow at the Alan Turing Institute, said: “Banks and other financial institutions have always been quick to adopt new technologies to make their operations more efficient, and the advent of LLMs is no exception. By bringing together experts from across the financial ecosystem, we were able to build a shared understanding of the use cases, risks, value and timelines for deploying these technologies at scale.”
Professor Lukas Spruch, Director of the Finance and Economics Programme at the Alan Turing Institute, said: “It is very positive to see the finance sector benefiting from the emergence of large-scale language models, and their deployment in this highly regulated sector could provide best practice for other sectors. This research shows the benefits of research and industry working together to assess the vast potential, as well as the practical and ethical challenges, of new technologies, to ensure they are deployed safely.”
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London, a comprehensive event taking place alongside other major events like BlockX, Digital Transformation Week and Cyber Security & Cloud Expo.
Find out about upcoming enterprise technology events and webinars hosted by TechForge here.
