The China Wealth Management 50-Person Forum will be held in Beijing in 2025. The theme of this year's conference is “Towards Building a Strong Financial Country in the 15th Five-Year Plan”. Leung Ding-pong, former chairman of the Hong Kong Securities Regulatory Commission, attended and delivered a speech. Liang Dingbang pointed out that artificial intelligence is growing rapidly at a rate of about 30% per year, the application coverage rate of large institutions has reached 85%, and costs have been drastically reduced and customer satisfaction has been greatly reduced. However, the in-depth application of AI in the financial sector faces four core risks of illusion, data bias, black boxes, and privacy compliance, which must be addressed carefully. He said that AI technology is currently progressing rapidly, mainly due to reduced hardware costs and increased computing power. However, AI is not limited to today's popular big language models; in the early days, it already covered various forms such as expert systems and machine learning. He cited AI-related achievements where both Nobel Prizes were awarded in 2024 as an example to explain their importance. However, AI applications, especially the “fill-in-the-blank” predictive reasoning used in big language models, have inherent flaws: first, it is easy to generate “illusions”, that is, to output inaccurate or fictional content; second, if training data is biased, the output results are also biased; third, the model decision-making process is complex and opaque, forming a “black box” that is difficult to explain; and fourth, there is a risk of improper use of data and infringement of personal privacy. These four major problems are the challenges that the financial industry must face to achieve “manageable risk.”

The China Wealth Management 50-Person Forum will be held in Beijing in 2025. The theme of this year's conference is “Towards Building a Strong Financial Country in the 15th Five-Year Plan”. Leung Ding-pong, former chairman of the Hong Kong Securities Regulatory Commission, attended and delivered a speech. Liang Dingbang pointed out that artificial intelligence is growing rapidly at a rate of about 30% per year, the application coverage rate of large institutions has reached 85%, and costs have been drastically reduced and customer satisfaction has been greatly reduced. However, the in-depth application of AI in the financial sector faces four core risks of illusion, data bias, black boxes, and privacy compliance, which must be addressed carefully. He said that AI technology is currently progressing rapidly, mainly due to reduced hardware costs and increased computing power. However, AI is not limited to today's popular big language models; in the early days, it already covered various forms such as expert systems and machine learning. He cited AI-related achievements where both Nobel Prizes were awarded in 2024 as an example to explain their importance. However, AI applications, especially the “fill-in-the-blank” predictive reasoning used in big language models, have inherent flaws: first, it is easy to generate “illusions”, that is, to output inaccurate or fictional content; second, if training data is biased, the output results are also biased; third, the model decision-making process is complex and opaque, forming a “black box” that is difficult to explain; and fourth, there is a risk of improper use of data and infringement of personal privacy. These four major problems are the challenges that the financial industry must face to achieve “manageable risk.”