By Polka Mishra

Imagine markets are falling sharply and you are worried about your portfolio. You ask a chatbot: “Should I sell everything now?”
Within seconds, it gives you a well-reasoned answer explaining why reducing risk could be sensible. It may cite economic uncertainty, weak sentiment and the potential for further losses.

Then you pause and ask a different question: “But should I stay invested for the long term instead?”

The chatbot easily gives you a persuasive case for doing exactly that – highlighting the importance of staying disciplined, avoiding market timing and allowing compounding to work.

AI can become a very sophisticated mirror – reflecting the emotion, bias or conclusion that the user has already formed. This can be helpful when it enables someone to explore competing views. But it becomes risky when an investor mistakes a convincing response for advice tailored to their actual circumstances.

Financial planning goes beyond spreadsheets and algorithms. A portfolio may look perfect on paper, but does it account for family commitments, career uncertainty, an upcoming property purchase, retirement needs, tax position and genuine tolerance for risk?

AI can consolidate financial information, forecast cash flow, run retirement scenarios, stress-test a portfolio and make complex concepts more accessible. But it cannot yet reliably place those outputs in the context of a client’s entire wealth journey – their priorities, trade-offs, changing circumstances and ability to stay invested when markets become uncomfortable.

The privacy risk is just as significant. The more personalised the AI’s response becomes, the more sensitive information a user may need to disclose. Income, assets, debts, portfolio holdings, family circumstances and retirement plans are not abstract data points – they are a detailed map of a person’s life.

That is your net worth being typed into a tool you may not control. In 2025, a researcher reportedly identified close to 100,000 ChatGPT conversations that users had voluntarily shared and that had become searchable through Google, including confidential contracts and private business information. Although OpenAI withdrew the relevant sharing feature, the episode was a reminder that convenience and confidentiality do not always move together.

Finally, there is the question of accountability. AI can generate an answer, but it does not bear the consequences if the information is inaccurate, incomplete or unsuitable. It is not a fiduciary, it does not know the full circumstances behind your decision, and there is no meaningful recourse when an unregulated chatbot gets it wrong. The responsibility for the outcome ultimately sits with the user.

Compare that with a licensed financial advisory firm. It operates within a regulatory framework, has conduct and suitability obligations, and must protect client information under the Personal Data Protection Act. More importantly, a professional adviser is accountable for the advice they provide.

The sensible approach is not to avoid AI. Use it to organise information, generate questions, compare scenarios and improve financial literacy. But when it comes to decisions involving your retirement, investments or family’s future, AI should inform your thinking, not replace your judgment.

AI belongs in the engine room. When it comes to your life savings, it should not be in the driver’s seat.