20家银行、14家技术供应商入围,香港金管局最新公布

Core Insights - The Hong Kong Monetary Authority (HKMA) and Hong Kong Cyberport announced the second phase of the Generative AI (GenAI) sandbox, marking a shift from exploring AI possibilities to promoting safe and reliable AI applications [1] - A total of 27 proposals from 20 banks and 14 technology providers were selected from over 60 submissions, evaluated based on innovation, technical complexity, and potential industry value [1][2] Group 1: Participants and Proposals - The 20 participating banks include Ant Bank (Hong Kong), Bank of China (Hong Kong), and HSBC among others [1] - The 14 technology providers include AIFT, Alibaba Cloud, and EY, with applications covering risk management, anti-fraud measures, and customer experience [2] - Selected proposals include AI-driven credit portfolio management, smart trade financing, and AI-generated investment insights [2] Group 2: Focus on AI Governance - The second phase of the sandbox emphasizes strengthening AI governance, with several proposals employing "AI against AI" strategies for automated quality detection of AI-generated content [2] - The sandbox serves as a testing ground for innovative defense mechanisms against increasing deepfake fraud risks, with participants conducting simulated attack-defense tests [2] Group 3: Case Study - PAObank - PAObank, in collaboration with OneConnect, is developing an anti-fraud strategy platform focusing on deepfake technology recognition, utilizing advanced AI facial recognition for real-time user verification [3] - The platform boasts a defense rate exceeding 99% against AI-generated images, having already been implemented in multiple financial institutions in mainland China [3] - PAObank's customer deposits surpassed HKD 8 billion as of Q3 this year, reflecting strong growth in personal banking services [4]