Core Insights - Responsible AI is transitioning from a practice of a few companies to an industry standard, with deeper collaboration emerging in the US-China AI ecosystem driven by the strategic foresight and practical actions of CIOs [1][9]. Group 1: CIO Role Evolution - The role of Chief Information Officers (CIOs) has evolved from traditional technology managers to core drivers of enterprise strategy, guardians of risk control, and bridges for cross-border technology cooperation [3]. - CIOs are now expected to balance innovation with risk management, requiring a shift in mindset to become strategic business enablers rather than just technical supporters [8][9]. Group 2: Responsible AI Adoption - Only 28% of US respondents view "responsible AI" as a core business priority, and only 33% have implemented clear applications across their organizations, indicating a significant gap in AI governance maturity [3][4]. - The rapid pace of AI technology evolution has outstripped the development of governance frameworks, leading to a low maturity level in responsible AI practices [4][10]. Group 3: Data Governance Importance - Data is recognized as the fuel for AI, with high-quality data being essential for generating valuable AI outcomes. Effective data governance is crucial for the successful implementation of responsible AI [7]. - Companies with established data governance frameworks see a 2.8 times higher success rate in AI projects compared to those without such frameworks [7]. Group 4: Global AI Regulation Perspectives - There are significant regional differences in AI regulation, with the US and China adopting a more relaxed approach compared to Europe and the Middle East, which favor stricter regulations [5]. - The EU's AI Act introduces stringent compliance requirements for high-risk AI systems, which can inhibit innovation, particularly for small and medium-sized enterprises [5]. Group 5: Multi-AI Model Strategy - A multi-AI model strategy is essential for global enterprises to navigate varying regulatory requirements and business needs across different regions [9]. - Companies must adapt their AI model choices based on local compliance and operational demands, ensuring flexibility in their AI deployments [9]. Group 6: Future of AI in Business - The future of AI is seen as a dual opportunity and challenge for CIOs, who must navigate technological advancements, regulatory differences, and data governance to drive responsible AI development [9]. - As AI technology continues to evolve, responsible AI practices are expected to become standard across industries, fostering deeper collaboration in the US-China AI ecosystem [9].

Group 1 - The role of Chief Information Officers (CIOs) has evolved from traditional technology managers to core drivers of enterprise strategy, guardians of risk control, and bridges for cross-border technology collaboration [2][3] - A recent PwC survey indicates that only 28% of U.S. respondents view "responsible AI" as a top business priority, and only 33% of companies have implemented clear applications across the organization [2][11] - McKinsey's 2024 global AI survey shows that while about 60% of companies have initiated AI projects, only 15% have established comprehensive AI governance frameworks, with average returns on AI investments falling short of the expected 30% [2][3] Group 2 - Responsible AI should not only focus on risk mitigation but also on helping businesses extract more commercial value from AI systems, transforming it from a compliance tool to a value extraction engine [3][4] - The low maturity of responsible AI practices is attributed to the imbalance between the rapid pace of technological iteration and the development of governance frameworks [3][4] - The emergence of AI agents has highlighted the inadequacy of traditional application management models, complicating the establishment of forward-looking governance frameworks [3][4] Group 3 - Global differences in AI regulation were discussed, with the U.S. and China seen as more relaxed compared to Europe and the Middle East, which adopt stricter regulatory approaches [4][5] - The EU AI Act categorizes AI systems by risk levels, imposing stringent compliance requirements on high-risk AI systems, which can inhibit innovation, particularly for small and medium enterprises [5][6] - A unified global AI standard is desired to reduce cross-border operational costs, similar to telecommunications standards [5][6] Group 4 - Data governance is crucial for responsible AI implementation, with high-quality data being essential for generating quality AI outcomes [6][7] - Companies must invest significant effort in data governance, ensuring proper data management and access control to prevent sensitive information leaks [6][7] - Organizations with established data governance frameworks see a 2.8 times higher success rate in AI projects compared to those without such frameworks [6][7] Group 5 - The evolution of the CIO role requires a balance of entrepreneurial spirit and a strong sense of responsibility, as they must drive innovation while safeguarding data security and compliance [7][8] - CIOs are now seen as strategic business enablers, leveraging core data assets to enhance productivity and differentiate business offerings [7][8] - The challenges posed by geopolitical uncertainties have led to a focus on "supply chain resilience" among global enterprises [7][8] Group 6 - The importance of a multi-AI model strategy was emphasized, as different AI models have varying service terms and usage restrictions, necessitating compliance with regional regulations [8][9] - CIOs must navigate the complexities of cross-border regulations while ensuring the selection of the most suitable AI models for their business needs [8][9] - The dual-supplier strategy is being adopted to mitigate risks associated with reliance on a single technology source [8][9] Group 7 - The rapid evolution of AI technology presents both opportunities and challenges for CIOs, who must adapt to changing landscapes and governance requirements [9][10] - The future of responsible AI is expected to shift from being a practice of a few companies to becoming an industry standard, driven by the strategic foresight and pragmatic actions of CIOs [9][10]

专题：2025年大湾区交易所科技大会 梁松光认为交易所需平衡创新与稳定，负责任 AI 五项最佳实践包括：第一，建立强有力的 AI 治理框 架；第二，促进人工监督，保持 "人机协作" 模式；第三，确保数据质量和多样性，减少偏见；第四， 提升全公司 AI 素养与风险意识；第五，通过本地或可信云基础设施安全受控地部署AI。 梁松光还表示，大湾区作为全球领先的科技创新集群，为交易所科技发展提供了独特优势。港交所期待 与深、广同业深化合作，在坚守市场诚信与稳定的前提下，充分释放 AI 在资本市场及更广域经济中的 赋能潜力，推动行业高质量发展。（新浪财经香港站） 责任编辑：郝欣煜 港交所集团首席资讯总监梁松光在"2025大湾区交易所科技大会"中表示，AI 或带来治理、隐私与安全 风险问题，建议交易所采用负责任AI的五项最佳实践，帮助利用AI创新业务和市场，同时不损害诚信 和稳定。 ...