来源：经济学家圈 2026年1月24日-25日，由香港中文大学（深圳）理工学院主办的 "翔龙鸣凤科学论坛：科学与哲学对话"在香港中文大学（深圳）举行。香港中文大学（深圳）校长、中国工程院院士徐扬生以"人工智能与东西方哲学思 想"为题进行主旨演讲。本文来源：徐杨生 徐扬生：人工智能与东西方哲学思想 根据徐扬生教授在2026年1月24日翔龙鸣凤科学论坛上的主旨演讲整理 第二，东西方哲学的侧重点很不同：西方哲学强调认识论、本体论多一点，而东方哲学则更偏向价值论，对认识论很少涉及，也不够严格，庄子是例外。 我为什么要先说以上这些？因为人工智能的发展，恰恰都会同时触及哲学的这三部分。 东西方哲学思想，我认为对人工智能是有指导作用的，我们正在构建的世界模型，这个模型怎么建？什么是智能？如何来评价智能？什么叫"好"的智能？ AI与人的边界在哪里？AI应该做人们心目中的"好孩子"吗？我们到底在追求什么？是在追求真理，还是在追求共识？这些问题不先弄清楚，AI的走向很 可能会不一样。 我们今天先从哲学开始讲起，粗略来说，哲学大致由三部分组成： 一是本体论：世界是什么； 二是认识论：我们怎么认识世界； 三是价值论/伦理学：我们如何 ...

Core Insights - The integration of artificial intelligence (AI) in the financial sector is seen as a critical opportunity for enhancing operational efficiency and service delivery, with a focus on addressing existing challenges in the industry [2][4][10]. Group 1: Current State of AI in Finance - Financial institutions are recognizing the necessity of adopting digital capabilities across various operational levels to navigate economic fluctuations [2]. - There is a consensus among financial entities regarding the importance of AI applications, although the pace and extent of implementation vary significantly [3]. - AI is primarily being utilized as an auxiliary tool in decision-making processes, with human oversight remaining crucial [3]. Group 2: Key Applications of AI - AI is being applied in several core areas, including digital marketing, risk management, and operational efficiency, with specific use cases such as automated portfolio management and enhanced customer profiling [5]. - The focus is on addressing pain points in financial services, such as improving transparency in technology finance and enhancing the matching of financial products to suitable clients [4][10]. Group 3: Challenges in AI Implementation - The uncertainty associated with AI technologies poses significant challenges, including potential risks in financial services due to computational errors [6]. - There are concerns regarding the clarity of responsibility between business and technical teams, as well as the difficulties in converting expert knowledge into AI training data [7]. - The banking sector faces five core challenges in AI deployment, including the need for optimized management systems and enhanced cross-departmental collaboration [7]. Group 4: Future Trends in AI in Finance - The service model in finance is expected to evolve towards a more seamless, less intrusive experience for customers, with ongoing transformations in physical channels [8]. - The financial sector will likely see a shift in human resource structures and an intensification of competition around data and open ecosystems [8]. - AI is anticipated to play a dual role as both a tool and a catalyst for theoretical innovation, necessitating a balance between technological advancement and ethical considerations [8]. Group 5: Recommendations for AI Development - Financial institutions are encouraged to enhance their technological maturity and create robust organizational frameworks to support AI integration [9]. - There is a call for collaboration between financial entities and external partners, such as academic institutions, to foster innovation in AI applications [9][10]. - Strengthening the infrastructure for AI applications, including improving credit assessment accuracy and establishing a secure data-sharing ecosystem, is essential for the future of finance [10].

当下，人人都对快速迭代的各类技术的巨大能量有了初体验，也深信其终将渗透并重塑经济和社会生活的各个领域，在金融行业亦不例 外。作为金融"五篇大文章"之一，金融机构尤其是银行对数字金融的重视和投入似乎怎么强调都不为过。 如今银行业的数字化进程到了哪一步？大模型等新兴技术强势来袭，银行业有怎样的思考和顾虑？新的数字化地基更新建成，下一步在技 术应用上会有怎样的突破？对此，21世纪经济报道《对话数字金融30人》高端访谈栏目近期专访了上海银行副行长、首席信息官胡德斌。 胡德斌拥有十分深厚的银行业数字化经历，拥有吉林大学软件工程博士学位。其职业生涯深度贯穿中国银行业信息化与数字化历程，曾历 任中国工商银行软件开发中心副总经理、数据中心（上海）副总经理等关键职务。自2016年出任上海银行副行长，并于2021年兼任首席信 息官以来，他主导推动了该行一系列重大科技战略工程。 近期，上海银行历时27个月的"智芯工程"圆满收官，新一代全栈信创核心系统成功投产。该工程不仅实现了从底层硬件到应用软件的全面 自主可控，更依托腾讯云TDSQL数据库与TCE专有云平台，完成了核心系统的平滑迁移与云化部署，标志着上海银行数字基础设施迈入了 全 ...

21世纪经济报道记者 方海平 上海报道 当下，人人都对快速迭代的各类技术的巨大能量有了初体验，也深信其终将渗透并重塑经济和社会生活 的各个领域，在金融行业亦不例外。作为"金融五篇大文章"之一，金融机构尤其是银行对数字金融的重 视和投入似乎怎么强调都不为过。 如今银行业的数字化进程到了哪一步？大模型等新兴技术强势来袭，银行业有怎样的思考和顾虑？新的 数字化地基更新建成，下一步在技术应用上会有怎样的突破？对此，21世纪经济报道《对话数字金融30 人》高端访谈栏目近期专访了上海银行（601229）副行长、首席信息官胡德斌。 胡德斌拥有十分深厚的银行业数字化经历，拥有吉林大学软件工程博士学位。其职业生涯深度贯穿中国 银行业信息化与数字化历程，曾历任中国工商银行软件开发中心副总经理、数据中心（上海）副总经理 等关键职务。自2016年出任上海银行副行长，并于2021年兼任首席信息官以来，他主导推动了该行一系 列重大科技战略工程。 近期，上海银行历时27个月的"智芯工程"圆满收官，新一代全栈信创核心系统成功投产。该工程不仅实 现了从底层硬件到应用软件的全面自主可控，更依托腾讯云TDSQL数据库与TCE专有云平台，完成了 核 ...

21世纪经济报道记者方海平上海报道 当下，人人都对快速迭代的各类技术的巨大能量有了初体验，也深信其终将渗透并重塑经济和社会生活 的各个领域，在金融行业亦不例外。作为"金融五篇大文章"之一，金融机构尤其是银行对数字金融的重 视和投入似乎怎么强调都不为过。 如今银行业的数字化进程到了哪一步？大模型等新兴技术强势来袭，银行业有怎样的思考和顾虑？新的 数字化地基更新建成，下一步在技术应用上会有怎样的突破？对此，21世纪经济报道《对话数字金融30 人》高端访谈栏目近期专访了上海银行副行长、首席信息官胡德斌。 胡德斌拥有十分深厚的银行业数字化经历，拥有吉林大学软件工程博士学位。其职业生涯深度贯穿中国 银行业信息化与数字化历程，曾历任中国工商银行软件开发中心副总经理、数据中心（上海）副总经理 等关键职务。自2016年出任上海银行副行长，并于2021年兼任首席信息官以来，他主导推动了该行一系 列重大科技战略工程。 近期，上海银行历时27个月的"智芯工程"圆满收官，新一代全栈信创核心系统成功投产。该工程不仅实 现了从底层硬件到应用软件的全面自主可控，更依托腾讯云TDSQL数据库与TCE专有云平台，完成了 核心系统的平滑迁移与云 ...

Core Insights - Peter Thiel is recognized as a unique figure in Silicon Valley, focusing on building foundational infrastructure rather than consumer products, driven by his understanding of "mimetic desire" and "creative monopolies" [1][2][4] Group 1: Palantir's Foundation and Philosophy - Palantir was founded by Thiel in 2003, aiming to address fundamental issues in the digital age rather than following trends in social applications [2][9] - The name "Palantir" is derived from a crystal ball in "The Lord of the Rings," symbolizing the creation of a digital mirror to understand and shape reality [4] - Thiel's philosophy emphasizes that companies should innovate fundamentally rather than compete in existing markets, leading to the establishment of a "creative monopoly" [2][4] Group 2: Investment Philosophy and Strategy - Thiel's investment strategy is characterized by building unique and irreplaceable value networks, as seen in his early investment in Facebook, which provided significant returns and strategic influence [12][14] - He focuses on long-term, high-risk projects that address fundamental problems, such as investments in biotechnology aimed at combating aging [14][24] - Thiel's approach contrasts with typical investors who chase short-term trends, as he seeks to create new possibilities rather than merely meeting existing demands [23][24] Group 3: Political Engagement and Influence - Thiel's support for Donald Trump in 2016 exemplifies his investment philosophy of positioning himself in undervalued areas, despite controversy in Silicon Valley [15][20] - His political investments aim to convert political capital into business advantages, enhancing his influence in both technology and governance [15][20] - Thiel's actions reflect a broader strategy of constructing a value network that spans technology, politics, and finance, aiming for a cohesive influence across sectors [20][21] Group 4: Philosophical Underpinnings and Future Vision - Thiel's investment decisions are informed by his philosophical beliefs, particularly the "power law," which suggests that a small number of key decisions yield the majority of results [17][18] - He seeks to redefine societal structures through his investments, aiming to address existential questions about life and technology [20][21] - Thiel's ultimate goal appears to be the reconstruction of world order based on his philosophical principles, challenging conventional norms and exploring the implications of technological advancements [20][21][24]

Core Viewpoint - Palantir Technologies has successfully transformed from a government contractor into a provider of AI infrastructure, leveraging a unique business model that combines complexity management and value personalization to create customized complex system solutions [5][55]. Group 1: Business Model Analysis - Palantir's business model is characterized by its ability to provide tailored solutions for complex problems, which distinguishes it from traditional software and consulting firms [8][15]. - The company has achieved a gross margin of 55% for scaled clients, with an average annual revenue of $10 million per client [7]. - Palantir's revenue is well-balanced between government and commercial sectors, with government revenue at $1.57 billion and commercial revenue at $1.3 billion [7]. Group 2: Historical Development and Key Milestones - Founded in 2003, Palantir initially focused on the government market, gaining significant trust and insights through early investments from the CIA's venture arm [21][22]. - The company began its commercial expansion in 2009 with a partnership with JPMorgan, marking a pivotal shift towards the commercial sector [24]. - In 2023, Palantir achieved its first annual profit of $217 million, with revenues reaching $2.225 billion, reflecting the success of its "Acquire-Expand-Scale" business model [28][30]. Group 3: Financial Model and Growth Mechanism - Palantir's financial strategy is based on a three-stage model: Acquire, Expand, and Scale, which emphasizes long-term investment over short-term profits [30][31]. - The company has diversified its revenue streams, successfully balancing government and commercial business, particularly after the launch of its AI platform [34]. Group 4: Competitive Advantages - Palantir's technological moat is driven by its ontology-based data integration capabilities, which create a "digital twin" of real-world objects and relationships [35][36]. - The Forward Deployed Engineers (FDE) model allows for deep customer engagement and rapid product iteration, enhancing customer relationships and service quality [37][38]. - The Apollo system supports the transition from consulting services to a scalable software company, enabling automated deployment and management of software [38]. Group 5: Market Position and Competitive Landscape - Palantir occupies a unique market position, often competing against clients' internal IT departments rather than traditional software vendors [39]. - The company's competitive advantages are sustainable, built on a combination of technology, data, relationships, and scale [41]. Group 6: Strategic Transformation in the AI Era - The launch of the AI Platform (AIP) marks Palantir's strategic shift into the AI era, integrating large language models with its existing data infrastructure [42][43]. - The financial performance post-AIP launch validates the effectiveness of this strategic transformation, with significant growth in commercial revenue [46].