Reading List: LLMs in Finance - UK Context#
The Alan Turing Institute and colleagues from HSBC and FCA recently launched a new report on the integration of LLMs in finance sector. This report was based on a comprehensive literature review on LLMs in banking and a consensus building workshop. Over forty attendees from major high-street banks, regulators, investment banks, insurers, payment services providers, as well as government and legal professionals working in financial services attended this workshop.
I was one of the leading researchers in this exciting study, where we explored opportunities and challenges in the integration of LLMs into financial services. You can access the report using this link: The Alan Turing Institute - LLMs in Finance Report.
Although the report explains the fundamentals in both LLMs and financial services, folks not familiar with the details might get lost while reading the report. So, I decided to list some reading materials that are key to understand the concepts discussed in this report. I suggest reading the materials in the presented order as each reading material is built upon the preceding materials.
Natural Language Processing by Elizabeth D. Liddy (2001): Although the LLMs are developed and popularised very recently (since ~2016), it is important to understand the early developments in the natural language processing (NLP) field. It presents an overall view on the developments in the NLP and AI between 1940-2000. It is a great summary that lists the key achivements in the NLP field.
Frontier AI: capabilities and risks – discussion paper (2023): Published before the AI Safety Summit - UK, the paper outlines the capabilities, risks, and cross-cutting challenges presented by the technology, pointing particularly to dangers around misuse, social harms, and loss of control.
University of Cambridge - Policy Brief: Generative AI (2024): The focus is assessing generative AI from an economic growth perspective. They answer three key questions: (1) What policy infrastructure and social capacity does the UK need to lead and manage deployment of responsible generative AI (over the long term)? (2) What national capability does the UK need for large-scale AI systems in the short- and medium-term? (3) What governance capacity does the UK need to deal with fast moving technologies, in which large uncertainties are a feature, not a bug?
Stanford HAI - Policy Brief: Safety Risks from Customizing Foundation Models via Fine-Tuning (2024): The document explains the safety risks focusing on custom fine-tuning of LLMs.
On the Societal Impact of Open Foundation Models by Kapoor et al. (2024): Analyzing the benefits and risks of foundation models with widely available weights.
The Alan Turing Institute - AI in Finance (2019): A literature review with the AI, ML and DL taxonomy as well as their various applications in the financial services industry.
Bank and FCA - ML in UK Financial Services (2019) Building upon a similar joint survey on 2019, this report findings representing near 300 firms, including banks, credit brokers, e-money institutions, financial market infrastructure firms, investment managers, insurers, non-bank lenders and principal trading firms.
Bank and FCA - AI Public-Private Forum (2022) It is the joint report from Bank of England (Bank) and the Financial Conduct Authority (FCA). They established the AI Public-Private Forum (AIPPF) in October 2020 to further the dialogue between the public sector, the private sector, and academia on AI. The aim of the AIPPF was to share information, deepen our collective understanding of the technology and explore how we can support the safe adoption of AI in financial services.