AI Financial Advisors: Trustworthy or Risky?

Artificial intelligence, particularly generative AI models like ChatGPT and Gemini, is rapidly entering the realm of personal financial management. A recent OECD report, 'AI and Personal Finance,' indicates that in 2025, 68% of South Korean adults aged 25-59 reported using generative AI for financial advice and tasks. This trend is observed globally, with surveys and research in countries like the U.S., UK, and Canada showing widespread adoption.

“

It is better to adjust the portfolio only when it deviates from the target allocation by a certain range, rather than adjusting it frequently. Set the stock portion to 40%, but only return to the original proportion when it deviates from the 35-45% range.

Consumers are leveraging AI for a variety of financial needs. This includes creating monthly budgets by inputting income and expenses, developing savings plans for goals like retirement or home purchases, and seeking advice on which debts to prioritize to minimize interest. AI is also used to find ways to improve credit scores. For investment decisions, users input factors like age, income, investment horizon, and risk tolerance to receive asset allocation recommendations. AI also acts as an educational tool, summarizing and explaining corporate earnings reports and economic news.

“

The importance of the proportion of domestic and foreign assets within the stock portfolio is more important than the current loss rate. Since U.S. stocks are defending against losses, let's maintain domestic stocks at only 25%. It's slightly below the target proportion, so buy 10 more shares and close the stock window.

Complex areas like insurance and taxes are also seeing significant AI utilization. A study across six countries, including Japan, the UK, and the U.S., found that 68% of consumers used AI for preliminary research before consulting an insurance agent, with 46% requesting comparisons of various insurance products. AI is also employed to understand tax filing procedures, clarify unfamiliar tax terms, identify applicable deductions, and check eligibility for welfare programs.

“

Generative AI can reduce information asymmetry in financial markets that is tilted to the disadvantage of consumers.

The OECD report highlights AI's potential to reduce information asymmetry in financial markets, benefiting consumers. The anonymity offered by AI allows users to comfortably ask sensitive questions about finances that they might hesitate to discuss with others. However, the report also issues a strong caution against blindly following AI's advice. The 'hallucination' phenomenon, where AI generates plausible but factually incorrect information or cites non-existent sources, is a significant concern in financial advice. Furthermore, AI's reliance solely on user-provided information limits its ability to grasp nuances like family structure, employment stability, or future expenses, potentially leading to unsuitable recommendations. Bias learned from training data can also manifest, as seen in instances where AI offered different advice based on gender, potentially disadvantaging certain groups.

“

AI cannot replace financial literacy. Users need to ask specific questions, verify the sources provided by AI, and compare important decisions with other data or expert opinions.