AI Tools for Financial Advice Show Practicality but Struggle with Vulnerability

Artificial intelligence (AI) tools such as ChatGPT, Claude, and Perplexity are increasingly being used for financial advice due to their ability to break down complex concepts quickly. However, recent research indicates that while these tools provide structured and practical advice, they have significant blind spots, particularly in addressing user vulnerabilities.

The study involved creating hypothetical scenarios to test how these AI models handle financial queries from users in vulnerable situations. Scenarios included a young graduate saving for a home, a pregnant woman planning for maternity leave, and a single parent considering cryptocurrency investment. Despite being given explicit details about the users' vulnerabilities, the AI models often failed to respond appropriately, sometimes offering advice that could exacerbate financial harm.

ChatGPT was found to be detailed and practical but did not recognize embedded vulnerabilities, relying solely on explicitly stated information. Perplexity was conservative, often suggesting professional financial advice, but lacked detail. Claude provided comprehensive recommendations but leaned towards self-guided planning, missing the nuances of user vulnerability.

The research highlights a critical issue: AI models are trained on large datasets that can reflect human biases, leading to assumptions based on stereotypes. For example, AI outputs often assumed traditional household financial roles, which did not match the explicit details provided in the scenarios.

Ultimately, the study concludes that while AI can be a powerful tool for financial brainstorming and fact-finding, it cannot yet replace the nuanced understanding and tailored advice that human financial advisors provide.