Xin Lang Cai Jing·2025-12-29 02:37Core Viewpoint - The 2025 Annual Meeting of the China Wealth Management 50 Forum emphasizes the theme of "Moving Towards a Financial Power during the 14th Five-Year Plan" and highlights the importance of artificial intelligence (AI) as a national strategy, with challenges and strategies for its application in the financial sector discussed by Yang Longru, General Manager of the Financial Technology Department of the Industrial and Commercial Bank of China (ICBC) [1][3][7]. Group 1: Challenges in AI Application - The development of foundational models (base models) is insufficient to meet the specific needs of the financial industry, as they perform well on general tasks but lack the specialized capabilities required for complex financial scenarios [3][8]. - There is a scarcity of high-quality financial datasets for training large models, with internal bank data being fragmented and inconsistent, making it difficult to convert expert knowledge into usable data for models [4][9]. - The transformation of business models from "+AI" to "AI+" presents significant challenges, as many AI applications remain siloed and do not effectively integrate across customer service or employee workflows, limiting the potential value of AI [4][8]. Group 2: Strategies for Overcoming Challenges - Emphasizing model training methods and enhancing adaptability through reinforcement learning and other post-training methods can improve the ability of large models to address specialized problems [5][10]. - Exploring knowledge engineering pathways to accelerate the conversion of high-quality knowledge into usable resources is essential, requiring the establishment of unified knowledge standards and effective integration of internal and external information [5][10]. - Embracing model transformation by actively exploring the application of enterprise super-intelligent agents can facilitate human-machine collaboration and break down departmental barriers, driving organizational change [5][10]. - Strengthening risk management and cautiously advancing direct customer services is crucial, necessitating increased resources for risk system development and the identification of both traditional and non-traditional security challenges [5][10].