Model-turned-broadcaster Byun Jung-soo revealed on the YouTube channel 'Bolbbangan Nyonnyonnyon' that she started stock investing due to artificial intelligence, purchasing SK Hynix shares at 1.75 million won per share based on ChatGPT recommendations and subsequently recovering her principal. This disclosure has sparked interest among investors as a growing number of retail traders in South Korea and globally turn to AI tools like ChatGPT, Gemini, and Claude for stock analysis and portfolio management. A March survey by Investing.com found that 62% of 938 US retail investors already use AI in investment decisions, while the Wall Street Journal reported in May that approximately 30% of investors on eToro employ AI for portfolio management, reflecting a broader shift toward AI-assisted trading strategies.
Byun shared her experience during a Bangkok trip with actors Bae Jong-ok and Yoon Hyun-sook, explaining that she now invests solely with profits after withdrawing her original capital. A 42-year-old homemaker identified as Ms. A also reported experimenting with AI stock recommendations, stating that while she does not expect outsized returns, the tools help her track market trends and company information. The practice has become sufficiently common that multiple books and online courses on AI-assisted investing have appeared in South Korea.
Global Surveys Confirm Widespread AI Adoption in Investing
Investing.com surveyed 938 US retail investors in March and found that 62% use AI in investment decisions, with 23.6% reporting regular usage. Among respondents, 26.6% stated they had made multiple trades based on AI-generated ideas. The Wall Street Journal cited eToro data in May indicating that approximately 30% of investors on the platform use AI for portfolio management. These figures align with anecdotal reports from South Korean investors who employ large language models to summarize news, screen stocks, and analyze financial statements.
Academic Research Reveals AI Limitations in Active Trading Strategies
A joint study by researchers from the University of Edinburgh, Sungkyunkwan University, and UCLA, published on the preprint site arXiv under the title "Can LLM-based Financial Investing Strategies Outperform the Market in Long Run?", tested AI-driven investment strategies across more than 100 stocks from 2004 to 2024. The research found that active strategies involving frequent AI-based rebalancing failed to outperform simple buy-and-hold approaches over the long term. The study attributed this underperformance to AI being overly cautious in bull markets and excessively aggressive in bear markets. However, the researchers noted that AI demonstrated clear utility in rapidly summarizing corporate disclosures, extracting key points from earnings reports, and organizing foundational data across multiple stocks.
A 35-year-old office worker identified as Mr. B reported discontinuing his paid AI subscriptions after encountering inconsistent analyses and occasional hallucinations, including incorrect current stock prices. He had used multiple AI platforms — ChatGPT, Gemini, and Claude — for cross-verification but found the discrepancies undermined his confidence in the tools.
Experts Warn Against Over-Reliance on AI for Buy-Sell Decisions
Professors from the University of Rome Tor Vergata and the University of Georgia published a paper in the June issue of the Journal of Financial Planning, cautioned that "AI responses sound confident but remain incomplete, potentially misleading, and subject to inaccuracy." The authors emphasized that delegating direct buy-sell timing decisions or position-sizing choices to AI carries significant risk. The Edinburgh-Sungkyunkwan-UCLA study similarly concluded that AI's value lies in information aggregation and preliminary screening rather than autonomous trade execution, particularly in high-volatility environments such as South Korea's KOSPI, which has experienced daily swings of 5–10% in recent periods.
FAQ
What stock did Byun Jung-soo buy based on ChatGPT recommendations?
Byun Jung-soo purchased SK Hynix shares at 1.75 million won per share following ChatGPT recommendations and has since recovered her principal, now investing only with profits.
How many US investors use AI for stock investing according to the March survey?
Investing.com's March survey of 938 US retail investors found that 62% use AI in investment decisions, with 23.6% using it regularly and 26.6% having made multiple trades based on AI-generated ideas.
What did the Edinburgh-Sungkyunkwan-UCLA study conclude about AI trading strategies?
The study, which tested AI-driven strategies across over 100 stocks from 2004 to 2024, found that frequent AI-based rebalancing underperformed buy-and-hold approaches long-term, though AI proved useful for summarizing disclosures and organizing company data.
