Three decades ago, Vasant Dhar was on the cutting edge of finance, building one of the earliest hedge funds powered by artificial intelligence. His work predated the current AI boom by years. Yet today, despite his deep expertise, he has no interest in handing over his personal investments to a chatbot like ChatGPT.

Dhar's reluctance isn't rooted in technophobia. As a professor at NYU and founder of the firm Socratic Technologies, he understands the power of machine learning better than most. He saw firsthand how AI could sift through massive datasets to uncover subtle patterns that human traders might miss. But he also knows the difference between a statistical model trained on clean historical data and a large language model that generates plausible-sounding but often flawed answers.

In Dhar's view, generative AI lacks the rigor and accountability required for serious money management. While it can summarize earnings reports or explain market jargon, ChatGPT cannot guarantee consistency or risk management. Investors should view these tools as research assistants, not portfolio managers. Dhar argues that real financial AI must be purpose-built, backtested, and transparent — qualities that black-box models like ChatGPT simply do not offer.

The lesson for today's investors: don't confuse conversational fluency with financial intelligence. AI has transformed Wall Street, but the technology that works for trading is not the same as the one that writes poetry. As Dhar puts it, trust the models that are built for the job — not the ones that sound like they know what they're doing.

Source: MarketWatch