At the "Global Capital Market Development Trends" forum, Zhang Jianhua, Director of the Financial Development and Regulatory Technology Research Center at Tsinghua University's PBC School of Finance, stated that the application of AI in the financial sector is not new, with uses such as small models and big data credit assessment emerging as early as two decades ago. Currently, the capabilities of large models have significantly improved, offering more substantial replacement of human roles in reasoning, integration, and collaboration, but they also introduce new challenges: first, insufficient understanding of the potential risks of large models; second, trial-and-error and waste in application-side investments, with large institutions pursuing multiple technical pathways simultaneously while small and medium-sized institutions face a vast investment gap.

Zhang Jianhua believes this investment disparity in AI could widen the AI divide, potentially leading to hundred-fold or thousand-fold differences in capability building between large, medium, and small Financial Institutions. This, in turn, could affect the structure of the financial system and the competitive landscape of the market, even triggering market exits for some institutions.

Zhang Jianhua emphasized that the financial industry's use of AI prioritizes safety, maturity, and controllability, not necessarily the most advanced top-tier models, with the key being suitability and feasibility. Models used for customer-facing applications must be subject to strict constraints, while internal empowerment can be relatively flexible; simultaneously, it is necessary to guard against impacts from external high-risk models.

Looking towards 2030, Zhang Jianhua believes the core keyword brought by AI is differentiation: there will be a divergence in capabilities between those who can and cannot use AI effectively, and for Financial Institutions, there will be a significant split between those that use AI well and those that do not.