What's Happening?
A new NYU Stern working paper, titled “Generative AI and Retail Investment” by finance PhD student Valerie Baldinger, indicates that retail investors using AI assistants are shifting their investment strategies. The study, which analyzed de-identified
data from Mezzi, a U.S. wealth management app, found that after adopting an AI assistant, investors increased their monthly share of index fund purchases by 7.6 percentage points and decreased single-stock purchases by 5.9 percentage points. This shift was observed among approximately 1,900 investors and 69,598 prompts. The AI assistant, built on OpenAI’s ChatGPT and Google’s Gemini models, often recommended index funds even when investors inquired about individual stocks. For instance, the Vanguard Total Stock Market ETF (VTI A) was the most frequently recommended ticker. This change in behavior led to a reduction in 'return loss,' a measure of expected return given up by holding diversifiable risk, by 43 basis points annually across the sample.
Why It's Important?
This shift in retail investor behavior, driven by AI recommendations, has significant implications for the U.S. financial markets and individual investors. Historically, efforts to encourage diversification and index fund adoption among retail investors have faced challenges, with low uptake rates for traditional advice. The study suggests that personalized advice from AI assistants is more effective, with investors acting on recommendations in 36.2% of cases where buy advice was given. This could lead to more diversified portfolios for retail investors, potentially reducing their risk exposure and improving long-term returns. Furthermore, the concentration of AI recommendations on a few specific funds, such as Vanguard’s VTI, could lead to a 'herding' effect, where a large number of retail investors flow into the same securities. This phenomenon could impact market dynamics and liquidity for these popular funds, potentially creating new challenges or opportunities for fund managers and market participants.
What's Next?
The findings suggest a growing influence of AI in personal finance and investment decisions. As AI wealth management tools become more sophisticated and widely adopted, there could be further shifts in how retail investors allocate their capital. Future developments might include more tailored AI advice, potentially leading to even greater diversification or, conversely, increased concentration in specific asset classes if AI models converge on similar recommendations. Regulators may also begin to examine the implications of AI-driven investment advice, particularly concerning potential market concentration and the responsibility of AI platforms. The study also highlights the impact of platform settings and instructions on AI recommendations, indicating that developers and wealth management apps have a significant role in shaping investor behavior through their AI configurations. Further research will likely explore the long-term effects of AI adoption on market efficiency and investor outcomes.
Beyond the Headlines
The emergence of AI as a significant influencer in retail investment decisions raises several deeper implications. Ethically, it brings to the forefront questions about the transparency and potential biases embedded within AI algorithms that guide investment advice. While the study indicates positive outcomes like reduced return loss, the 'black box' nature of some AI models could obscure the rationale behind recommendations, making it difficult for investors to fully understand the risks. Legally, the role of AI in providing financial advice may necessitate new regulatory frameworks to ensure investor protection and accountability. Culturally, the increasing reliance on AI for financial decisions could alter the traditional relationship between investors and human financial advisors, potentially leading to a greater democratization of sophisticated investment strategies but also a reduced understanding of underlying financial principles. The concentration of recommendations on a few funds also hints at potential systemic risks if a large segment of the market moves in unison based on AI guidance, which could amplify market volatility during downturns.













