2025年，全球智能化浪潮风起云涌，人工智能领域创新呈群体性突破之势，语言大模型、多模态大模型 和具身智能等领域日新月异，推动人工智能向更高效、强智能的方向快速发展。同时，人工智能带来的 风险挑战引发广泛关注。国际社会需进一步凝聚共识，确保人工智能为人类所利用、为人类所掌控，朝 着向善普惠的方向发展，成为造福人类的国际公共产品。本期邀请4位国内外专家，就加强人工智能全 球治理分享真知灼见。 薛澜（清华大学人工智能国际治理研究院院长） 托马斯·葛格里（匈牙利科学院数学研究所学部委员、布达佩斯技术和经济大学荣誉教授） 安东尼·穆恩萨米（南非豪登省合作治理部首席主任） 英厄·冯贝勒（光辉国际中东公司高级客户合伙人） 治理赤字 国际公共产品。中国将善治逻辑引入人工智能领域，聚焦推动人类文明整体进步，代表了构建包容、健 康、可持续的人工智能全球治理体系的正确方向。 以人为本 只有把"人"放在核心位置，人工智能技术才能真正推动社会进步 托马斯·葛格里：面对层出不穷的挑战，如何对人工智能主导的信息空间进行系统性治理成为重要议 题。各国研究者普遍认为，需要建立兼具本国和全球层面的人工智能治理框架，涵盖理念、伦理、法律 等各方 ...

AI是否存在泡沫？对AI技术的监管该严还是松？如何确保这项技术发展始终服务于人的福祉？这些有 关AI的核心争论，或许能从"上善AI"的理念找到解答方向。 12月18日下午，南方都市报、南都数字经济治理研究中心在北京举办"第九届啄木鸟数据治理论坛"，主 题聚焦"AI安全边界：技术、信任与治理新秩序"。十四届全国政协委员、工业和信息化部原副部长王江 平以《上善AI：以对齐求善治》为题发表主旨演讲，他引用老子"上善若水"的理念，尝试勾勒中国AI治 理愿景。 12月18日，南都在京举办第九届啄木鸟数据治理论坛。 王江平认为，当前国际AI治理理念分歧较大，阵营化趋势明显，迫切需要一种东方古老智慧统一治理 愿景。国内AI监管同样需要一套科学、敏捷的框架体系，来平衡AI发展其间的激进与焦虑。 AI系统向"智能实体"转变，但治理领域进展有限 随着AI技术的演进和产业应用，AI治理问题日益显现。从AI生成"流浪汉入家"到AI仿冒名人带货，从 AI换脸拟声诈骗到未成年人过度沉迷聊天机器人……这些案例显示，当AI技术能力不断提升，安全边 界却在模糊。 王江平在演讲中提到，大模型加速落地的同时，由模型幻觉所引发的内容风险愈发成为热点 ...

Group 1 - The use of AI does not necessarily lead to a decline in intelligence; this question is overly simplistic and reminiscent of outdated industrial-era concerns [1][7] - Current AI systems primarily absorb human knowledge, functioning similarly to a talking encyclopedia, but they lack the ability to interpret non-verbal cues and emotional contexts [3][4] - AI's learning is based on vast amounts of data, yet the underlying values and contexts of this data remain difficult to assess, raising concerns about the potential biases in AI outputs [4][8] Group 2 - The importance of companionship in human development suggests that private AI applications, such as AI social companions and toys, could represent a significant market opportunity [2][9] - The evolution of education emphasizes the need for cognitive education, which is crucial in the AI era, as it shapes how individuals perceive and interact with the world [9] - The disparity in AI usage can exacerbate existing knowledge gaps, highlighting the need for effective AI tools to bridge these divides [12]

Core Viewpoints - The question of whether AI leads to a decline in intelligence is not a binary issue and reflects a misunderstanding similar to questions from the industrial era [3][10] - Human cognition is still in its early stages of understanding, with human thought characterized by leaps and sudden changes that are not yet fully explained [3][8] - Current AI systems primarily absorb human knowledge, functioning more like a talking encyclopedia, but they lack the ability to interpret non-verbal cues and emotional contexts [6][8] Group 1: AI and Human Cognition - AI's learning is based on vast amounts of human-generated data, but the implications of the background and values of this data remain uncertain [4][12] - The interaction with AI should be seen as a collaborative process that enhances human thinking rather than a simple tool for information retrieval [11][15] - The importance of questioning and challenging AI outputs is emphasized as a means to foster deeper cognitive engagement [11][12] Group 2: The Role of AI in Education and Development - The development of foundational skills such as language, logic, and cognitive abilities is increasingly important in the AI era [13][14] - The concept of "companionship" in human development is paralleled in the potential market for private AI applications, such as AI companions and toys [4][14] - Educational approaches should shift towards cognitive enhancement rather than mere knowledge transmission, encouraging discussions with AI to deepen understanding [14][15] Group 3: The Digital Divide and Social Diversity - The emergence of AI has the potential to equalize knowledge access, but disparities in AI usage can widen the gap between different user groups [16] - The notion of an "intelligence gap" may stem from a misperception of one's position in society, highlighting the need for diverse perspectives [16] - The subjective experience of life and happiness varies greatly among individuals, underscoring the importance of embracing social diversity [16]

Core Viewpoint - The Chinese government has proposed the establishment of a World Artificial Intelligence Cooperation Organization to address the rapid evolution of AI technology and its implications for global governance and safety [1][2][3] Group 1: Strategic Significance - The initiative aims to enhance technological collaboration and innovation while contributing to a global AI governance framework and addressing safety risks systematically [1] - The organization is seen as a necessary response to the dual-edged nature of AI technology, promoting global consensus on technological development and bridging the "intelligence gap" [1][2] Group 2: Global Impact and Cooperation - AI's borderless nature means that breakthroughs or failures in one country can have global repercussions, necessitating international cooperation [2] - The establishment of the organization will facilitate policy coordination among governments, experience sharing among enterprises, and collaborative research efforts, fostering a new paradigm of "co-governance and shared benefits" [2][3] Group 3: Addressing the "Intelligence Gap" - The organization aims to shift technology development from "monopolistic enjoyment" to "common prosperity," promoting equitable access to AI technologies [3] - It will work towards creating mechanisms for technology sharing, establishing fair rule-making platforms, and building AI capacity in developing countries, transforming them from mere "receivers" to active "participants" [3]

Core Viewpoint - The article emphasizes the urgent need to bridge the digital divide exacerbated by the rapid development of artificial intelligence and digital technologies, which significantly impacts global economic growth and social equity [1][2][9]. Group 1: Digital Divide - The digital divide manifests in three dimensions: access, usage, and capability, leading to a more pronounced "intelligent divide" as AI technology evolves [1][2]. - By the end of 2024, nearly 2.6 billion people globally will still lack internet access, representing one-third of the world's population, with only 35% of the least developed countries having internet users [2]. - The disparity in digital technology usage is evident, with high-income countries having 93% internet usage compared to only 39% in landlocked developing countries [2]. Group 2: International Cooperation - International organizations are enhancing collaborative mechanisms to address the challenges posed by the digital divide, with major countries implementing digital strategies to improve infrastructure and digital literacy in underserved areas [3]. - China is actively participating in global digital governance, promoting infrastructure development in partner countries through initiatives like the Belt and Road Initiative, significantly improving network coverage and digital capabilities in Africa [3]. Group 3: AI Challenges and Opportunities - The development of AI presents both challenges and opportunities in bridging the digital and intelligent divides, with issues such as data concentration and regulatory discrepancies hindering progress [5][6]. - AI technology can enhance resource sharing and digital literacy, with innovations like lightweight models enabling effective applications in resource-limited areas [6]. Group 4: Collaborative Efforts - Bridging the digital and intelligent divides requires a concerted effort from governments, businesses, and international organizations to create an inclusive AI ecosystem and enhance digital literacy [7][8]. - There is a call for open resource sharing and collaboration to develop adaptable AI technologies that cater to the needs of developing regions [7][8].

Core Points - The Chinese government emphasizes that digital dividends should not lead to digital hegemony and that the intelligent revolution should not widen the digital divide [1] Group 1 - The second phase of the Artificial Intelligence Capacity Building Seminar commenced in Beijing on May 12, with participation from nearly 40 countries and international organizations [1] - This seminar is part of the implementation of the Global Artificial Intelligence Governance Initiative and aims to strengthen international cooperation in AI capacity building [1] - China advocates for fair and inclusive global development in artificial intelligence and aims to be a promoter and pioneer in international cooperation on AI capacity building [1] Group 2 - The Chinese government is committed to the principles of consultation, contribution, and shared benefits, aiming to enhance communication and cooperation with various parties [1] - The goal is to provide public goods to the international community and share the benefits of AI with developing countries [1] - The focus is on ensuring that artificial intelligence better serves global development [1]