Generative AI in Finance

Generative AI uses large language models and similar technologies, trained on vast amounts of data, to produce human-like text and other outputs in response to prompts. Its use in finance grew rapidly after such tools became widely available in late 2022.

Common applications include customer service chatbots and virtual assistants; summarising research reports, earnings calls and regulatory documents; drafting reports, emails and client communications; helping analysts and advisers find information quickly; assisting with coding and software development; and supporting compliance, such as reviewing documents and monitoring communications.

For example, a bank might give its financial advisers an AI assistant that searches internal research and drafts summaries for client meetings, while the adviser checks and approves the final content.

Generative AI also brings risks. Models can produce confident but incorrect information, sometimes called hallucinations; they may reflect biases in their training data; and they raise concerns about data privacy, security, intellectual property and accountability.

Regulators expect firms to manage these risks under existing rules, and new rules, such as the EU AI Act, set additional requirements. Firms typically keep humans involved in important decisions.