The "Caijing" Annual Conference 2026: Forecast and Strategy & 2025 Global Wealth Management Forum was held in Beijing from December 18 to 20, 2025. Liu Gang, Managing Director of Research at CICC and Chief Overseas Strategist, shared three core perspectives on the AI bubble debate:
1. The term "bubble" should be viewed as a neutral concept rather than inherently negative. 2. Obsessing over whether AI will become a bubble is less meaningful, as the bubble formation phase often coincides with the market's strongest rally. Exiting too early may mean missing opportunities. 3. The key to evaluating a bubble lies in whether "investment aligns with demand" and whether "investment exceeds one's capacity."
Liu revealed that AI has profoundly transformed research methodologies. For instance, interpreting the Central Economic Work Conference can now be accelerated using AI to compare historical meeting content and generate analytical reports, significantly boosting efficiency. "Everyone can be their own analyst," he noted.
However, Liu emphasized three critical limitations preventing AI from fully replacing human researchers: 1. AI suffers from "hallucination" issues, and users must possess basic domain knowledge to discern the accuracy of AI-generated answers—especially for specialized queries like theoretical physics. 2. Knowledge differs from wisdom. While AI provides information, the key to problem-solving lies in human questioning techniques and decision-making logic. "Guiding AI on how to proceed" matters more than merely obtaining answers. 3. Financial markets are heavily influenced by human emotions. AI's binary computations cannot precisely capture market fluctuations driven by sentiment—a core strength of human researchers.
Liu concluded that AI's impact on research and daily life is irreversible. Investors and researchers must proactively adapt while recognizing technological boundaries, leveraging human strengths in cognitive judgment, emotional perception, and strategic decision-making.
Comments