Peter Dillard, Global Head of Investment Engineering at Dimensional Fund Advisors (DFA), stated on May 28 (local time) that artificial intelligence can help predict market volatility but cannot reduce volatility itself, during a webinar hosted by the World Economic Research Institute. The remarks came as Korean stocks triggered circuit breakers, with the KOSPI falling over 10% and the KOSDAQ dropping over 8% following a sharp decline in US semiconductor stocks. DFA, which manages $1 trillion in assets, organized the webinar titled 'Portfolio Investment Strategies in the AI Era' to examine AI's actual applications and limitations in asset management amid growing market interest in AI-driven investment approaches.
Dillard States AI Predicts Volatility But Does Not Stabilize Markets
Dillard addressed whether AI could stabilize volatile markets such as Korean stocks. He stated that AI can assist in predicting periods or patterns of increased market volatility, but clarified this differs from reducing volatility itself. "Financial markets contain extensive information related to volatility, including prices and options," Dillard explained. "AI is useful for collecting and analyzing this information more quickly."
He emphasized that faster information reflection does not equate to market stabilization. "As new information is reflected in the market more quickly, price fluctuations can also increase," Dillard stated. "AI can make information reflect faster, but it does not make the market more stable." He added that circuit breakers serve as mechanisms allowing investors to view the market calmly for a moment, while AI does not play a role in reducing market volatility itself.
DFA Executive Expresses Caution on AI Supercycle Duration
Dillard expressed a cautious stance on the AI supercycle, a major market theme. "It is difficult to say how long the AI supercycle will last," he stated. "This phenomenon is occurring because demand currently far exceeds supply." He explained that if demand continues, companies will engage in long-term infrastructure investment and production expansion, but added that "nothing lasts forever."
Regarding AI's future role, Dillard cautioned against viewing AI as omnipotent. "I do not believe AI will eliminate all jobs or lead to mass layoffs," he stated. "AI will ultimately become a tool and will be utilized across individuals' lives and work." He emphasized that "what we need to do now is learn, research, and explore AI," adding that "answers to most matters lie somewhere in the middle rather than at either extreme."
DFA Tests AI Applications in Asset Management with Mixed Results
Dillard presented cases where DFA applied AI to actual asset management operations. He reported that AI demonstrated high efficiency in coding support and summarizing and analyzing complex corporate activity documents. However, AI's cost-effectiveness fell short of expectations in tasks such as anomaly detection in data and extracting key information from extensive bond covenants.
Dillard proposed establishing an "evaluation framework" to determine AI utilization based on task characteristics rather than unconditional adoption. He suggested first examining whether a task requires precise answers, whether draft creation alone suffices, whether experts can quickly verify results, and whether errors can be easily reversed.
Dillard Emphasizes Human Responsibility Over AI in Final Investment Decisions
Dillard offered a cautious view on whether AI can enhance investment performance. "Even if AI quickly synthesizes vast information and improves prediction accuracy, portfolio performance does not always improve," he stated. "Market prices reflect the information and judgment of numerous investors, making it difficult for AI to consistently beat this collective intelligence."
He diagnosed that "if universal data and universal AI are utilized, competitors are also likely to use similar tools," adding that "it is difficult to secure sustained competitive advantage with universal AI alone." Dillard emphasized that "asset managers are entrusted not simply with tasks from clients but with responsibility for problem-solving," stating that "final investment decisions and the resulting responsibility must be borne by people, not AI."
FAQ
What did Peter Dillard say about AI and market volatility on May 28?
Peter Dillard, Global Head of Investment Engineering at DFA, stated during a webinar on May 28 (local time) that AI can help predict market volatility patterns but cannot reduce market volatility itself. He explained that AI enables faster information collection and analysis, but faster information reflection can increase price fluctuations rather than stabilize markets.
How did DFA's AI applications perform in asset management tasks?
Dillard reported that DFA's AI applications showed high efficiency in coding support and document summarization but delivered cost-effectiveness below expectations in data anomaly detection and extracting key information from complex bond covenants. He proposed using an evaluation framework to assess AI suitability for specific tasks rather than adopting AI universally.
Why did Dillard express caution about AI's role in investment performance?
Dillard stated that AI's ability to synthesize information quickly and improve prediction accuracy does not guarantee better portfolio performance. He explained that market prices already reflect collective intelligence from numerous investors, making it difficult for AI to consistently outperform. He emphasized that final investment responsibility must remain with people, not AI, as asset managers bear accountability for problem-solving beyond simple task execution.
