Liu Jun, President of Industrial and Commercial Bank of China (ICBC), stated at the Lujiazui Forum in Shanghai that as AI reshapes the financial industry, commercial banks face a more "painful" transformation than central banks in AI deployment, especially when deploying AI tools at scale, as commercial banks bear direct financial and legal responsibilities. He pointed out that central banks primarily focus on regulation and policy, whereas if AI tools cause problems, commercial banks must compensate victims and provide final answers to the market. Liu Jun also mentioned four mismatches between AI and traditional banking in terms of disclosure, data storage, sovereignty, and credibility, and stated that ICBC has deployed hundreds of AI models and introduced a "proof of benefit" framework to mitigate operational risks.

Eddie Yue, Deputy Chief Executive of the Hong Kong Monetary Authority (HKMA), also expressed concerns about liability attribution when banks adopt agent-based AI. Last week, the HKMA unveiled an updated AI governance framework, focusing on accountability, resilience, and trust, aiming to create an environment where financial institutions can safely "fail fast," with regulators being more facilitative than supervisory.