科协要闻丨全球人工智能治理科技社团倡议

转自 中国科协之声

人工智能作为引领新一轮科技革命和产业变革的战略性技术，正深刻重塑世界经济格局、社会治理模式与人类生活方式。我国在人工智能全球治理中始终坚持发展与安全并重、创新与规范协同，提出《全球人工智能治理倡议》，倡导以人为本、智能向善，推动建立开放、公正、包容、有效的全球人工智能治理体系。

为凝聚全球学术共识，协力共同促进人工智能治理，在中国科协组织下，中国人工智能学会等16家单位，共同起草了《全球人工智能治理科技社团倡议》，紧密围绕确保有益、赋能民生，恪守安全、防范风险等内容发出倡议。一起看→

全球人工智能治理科技社团倡议

人工智能作为引领新一轮科技革命和产业变革的战略性技术，正深刻重塑世界经济格局、社会治理模式与人类生活方式。当前，全球人工智能技术快速发展，既为经济社会发展带来重大机遇，也带来一系列全球性治理新挑战，关乎全人类共同福祉与未来发展走向。中国在人工智能全球治理中始终坚持发展与安全并重、创新与规范协同，提出《全球人工智能治理倡议》，倡导以人为本、智能向善，推动建立开放、公正、包容、有效的全球人工智能治理体系。科技社团作为人工智能研究创新、理念传播、实践推动的核心力量，肩负着凝聚治理共识、筑牢治理根基、赋能治理实践的重要使命。为凝聚全球学术共识，协力共同促进人工智能治理，发起以下倡议：

确保有益、赋能民生

以增进人类共同福祉为人工智能学术研究的根本遵循，确保技术发展始终服务全人类共同利益。恪守全人类共同价值，尊重各国历史文化、社会制度和发展道路差异，引导人工智能在应对全球性发展难题和挑战中发挥积极作用。鼓励开展实践导向研究，总结推广可复制的技术惠民案例，促进技术成果和有益经验合理共享，使人工智能真正成为增进全人类福祉、助力可持续发展的重要力量。

恪守安全、防范风险

将安全作为人工智能学术研究与治理的底线要求，确保技术发展始终处于人类可控范围。坚守伦理先行，将科技伦理规范嵌入人工智能研究全生命周期，建立健全学术伦理审查机制，推动伦理审查标准化、制度化、常态化，筑牢伦理与安全防线。联合开展安全治理前沿研究，确保数据来源合法和安全，尊重知识产权、个人信息权益，重点关注算法滥用、虚假信息扩散、隐私泄露、模型操控等现实风险，以及人工智能失控、自我复制、自主逃逸等潜在系统性风险，探索风险研判、监测预警和协同防控体系，推动技术创新与治理规则协同演进。

秉持公平、促进均衡

尊重各国平等参与人工智能研究与治理的权利，以及其基于自身国情制定的人工智能治理路径与政策，反对技术霸权、学术壁垒、排他性“小圈子”及不合理技术垄断，推动数据获取、人才培养、规则制定等环节更好体现发展中国家与弱势群体诉求。加强与发展中国家的学术交流与能力建设，通过全球联合研究、联合培养、平台共建等方式，助力其提升人工智能研究与治理能力，共同弥合全球智能鸿沟，营造公正包容、开放共享的全球学术生态。

交流互鉴、携手合作

建立跨界多元协同交流机制，打破创新合作壁垒，凝聚各国科学家智慧，携手应对共同挑战，推动全球人工智能治理知识体系交流互鉴与建设。立足实践需求，对接各国学术机构、政府部门、国际组织等多方主体，关注合作诉求，寻求“最大公约数”。积极参与全球人工智能风险治理、伦理规范和标准互认，推进在联合国框架下成立国际人工智能治理机构的进程，推动学术共识向政策规则与治理实践转化。坚持以科学研究和事实证据支撑人工智能治理，不断提升治理规则的科学性、合理性、有效性和公正性。

向善传播、凝聚共识

多方式、多渠道开展人工智能科普与对话，引导公众正确认知技术发展阶段与潜在风险，提升全民科学素养，回应社会关切。完善学术出版审核机制与规范，强化学术出版伦理把关，在尊重知识产权和数据安全的前提下推进成果开放共享，打造权责清晰、可审核、可监督、可追溯、可信赖的学术诚信与成果传播体系，为全球人工智能治理提供坚实知识支撑。高度重视人工智能对科学研究范式带来的深刻变革，围绕科研伦理、学术诚信和科学文化等重要议题加强研讨，凝聚治理共识，引领数智时代科学研究健康发展。

人工智能发展跨越国界，治理责任需要全球共担。科技社团应坚守学术使命与职业操守，强化责任担当、协同行动，以开放包容的姿态加强国际学术交流合作，以务实有效的举措推动智能向善，为促进人工智能技术造福于全人类，推动构建人类命运共同体贡献智慧和力量。

Initiative of the Global Science and Technology Society

on AI Governance

As a strategic technology driving the latest round of scientific and technological revolution and industrial transformation, artificial intelligence (AI) is profoundly reshaping the global economic landscape, social governance models, and human lifestyles. Currently, the fast development of global AI technology has brought huge opportunities for economic and social development, but also a series of new global governance challenges, which bear on the wellbeing of humanity and its future development trajectory. In global AI governance, China consistently adheres to putting equal emphasis on development and security, and harmonizing innovation with regulation, and has proposed the Global AI Governance Initiative. It advocates a people-centered approach and developing “AI for good”, promoting the establishment of an open, fair, inclusive, and effective global AI governance system. As the core force driving AI research innovation, the dissemination of ideas, and the advancement of practical applications, the science and technology society shoulders the vital mission of building governance consensus, solidifying its foundations, and empowering its practices. To build global academic consensus and work together to jointly promote the governance of AI, we hereby launch the following initiatives:

Ensuring Beneficial Outcomes and Empowering Livelihoods

Taking the goal of increasing the wellbeing of humanity as the fundamental guiding principle for AI academic research, we must ensure that technological development consistently serves the common interests of humanity. We should abide by humanity’s common values, respect the differences in histories, cultures, social systems, and development paths of various countries, and guide AI to play a positive role in addressing global development problems and challenges. We encourage practice-oriented research to synthesize and promote replicable cases of technology benefiting the public, facilitate the equitable sharing of technological achievements and beneficial experiences, and make AI a truly vital force for increasing the wellbeing of humanity and supporting sustainable development.

Upholding Security and Preventing Risks

Establishing security as the baseline requirement for AI academic research and governance, we must ensure that AI always remains under human control. Adhering to the principle of putting “ethics first”, we should embed scientific and technological ethical norms throughout the entire life cycle of AI research. We aim to establish and improve academic ethical review mechanisms, promote the standardization, institutionalization, and normalization of ethical reviews, and build a robust defense line of ethics and security. We should collaboratively conduct frontier research on security governance, ensuring the legality and security of data sources, and respecting intellectual property rights and personal information rights and interests. We should focus on immediate risks such as algorithmic abuse, the spreading of disinformation, privacy leakage, and model manipulation, as well as potential systemic risks like AI loss of control, self-replication, and autonomous escape. Furthermore, we should explore systems for risk assessment, monitoring, early warning, and collaborative prevention and control, driving the synchronized evolution of technological innovation and governance rules.

Upholding Fairness and Promoting Balanced Development

We respect the right of all countries to participate equally in AI research and governance, as well as their AI governance paths and policies formulated based on their respective national conditions. We strictly oppose technological hegemony, academic barriers, exclusive “small circles”, and unreasonable technological monopolies, striving to better reflect the demands of developing countries and vulnerable groups in areas such as data access, talent cultivation, and rule-making. We should strengthen academic exchanges and capacity building with developing countries, assisting them in enhancing their AI research and governance capabilities through global joint research, collaborative training, and shared platforms. Together, we aim to bridge the global gap in AI and foster a fair, inclusive, open, and shared global academic ecosystem.

Fostering Exchange, Mutual Learning, and Collaborative Cooperation

We should establish cross-sectoral and multi-stakeholder collaborative exchange mechanisms, dismantle barriers to innovation and cooperation, and pool the wisdom of scientists worldwide to jointly address shared challenges, thereby promoting the exchange, mutual learning, and construction of a global AI governance knowledge system. Grounded in practical needs, we should engage with diverse stakeholders, including academic institutions, government bodies, and international organizations. We should attentively address cooperation demands and seek the “broadest common ground”. We should actively participate in the global governance of AI risks, ethical norms, and the mutual recognition of standards, advance the process to establish an international institution to govern AI within the United Nations framework, and promote the translation of academic consensus into policy rules and governance practices. We persist in supporting AI governance with scientific research and factual evidence, continuously enhancing the scientific rigor, rationality, effectiveness, and fairness of governance rules.

Promoting “AI for Good” Communication and BuildingConsensus

We should conduct AI science popularization and public dialogue through diverse channels, guiding the public to accurately understand the current stages of technological development and its potential risks, enhancing overall scientific literacy, and responding to societal concerns. We should refine academic publishing review mechanisms and norms, and strengthen ethical oversight in academic publishing. While respecting intellectual property rights and data security, we should promote the open sharing of research findings, and build a trustworthy academic integrity and achievement dissemination system with clear rights and responsibilities that can be reviewed, monitored, and traced, thereby providing robust intellectual support for global AI governance. We attach great importance to the profound changes AI brings to scientific research paradigms, and will strengthen discussions on critical issues such as research ethics, academic integrity, and scientific culture, to build governance consensus and guide the healthy development of scientific research in the era of digital intelligence.

The development of AI transcends national borders, and the responsibility for its governance must be shared globally. The science and technology society should uphold its academic mission and professional ethics, strengthen its sense of responsibility, and take coordinated action. With an open and inclusive posture, we should enhance international academic exchange and cooperation, advance the development of “AI for good” through pragmatic and effective measures, and contribute our wisdom and strength to promoting AI technologies to benefit humanity and building a community with a shared future for mankind.

*16家单位包括中国自动化学会、中国仪器仪表学会、中国图学学会、中国电子学会、中国计算机学会、中国通信学会、中国仿真学会、中国图象图形学学会、中国人工智能学会、中国指挥与控制学会、中国密码学会、中国科学学与科技政策研究会、中国科学技术期刊编辑学会、国际数字地球协会、世界机器人合作组织、国际智能制造联盟。

责　　编：胡安妮

审　　核：张敬一

复　　核：刘兴平

值班编委：谭华霖