**中外顶尖专家在京签署《北京AI安全国际共识》：划清AI风险红线，禁止自我复制与生化武器应用**
据腾讯科技报道，一场由智源研究院发起的重要会议在北京落幕，会上，包括图灵奖得主约书亚·本吉奥、杰弗里·辛顿、姚期智等在内的数十位国际知名专家共同签署了一项具有里程碑意义的文件——《北京AI安全国际共识》。这一共识旨在为人工智能的发展划定安全边界，保障全球社会的长远利益。
《共识》明确提出了人工智能的“风险红线”，将“自主复制”和“应用于生化武器”列为明确禁止的行为。专家们担忧，若AI具备自我复制能力，可能会导致无法控制的扩散风险；而将其应用于生化武器，则将极大地威胁人类的生命安全和国际和平。
此外，“权力寻求”、“协助不良行为者”和“欺骗”也被列为AI发展的禁区。这些规定旨在防止AI技术被滥用，以维护全球网络安全和社会秩序。《共识》的出台，反映了国际社会对AI伦理和安全问题的高度重视，为全球AI领域的健康发展提供了指导原则。
智源研究院作为此次活动的发起者，展现了其在人工智能研究与治理领域的领导力。该共识的签署，不仅是对AI技术可能带来的潜在风险的警示，也是对全球科研界和产业界的一份责任书，敦促各方在推进技术创新的同时，要兼顾道德和法律的约束，确保AI技术的可控和有益。
未来，随着AI技术的深入发展，如何在保障创新与安全之间找到平衡，将是各国政府、科研机构和企业共同面临的挑战。《北京AI安全国际共识》的签署，标志着全球对此问题的讨论迈出了实质性的一步。
英语如下：
**News Title:** “Global Top Experts Reach Consensus: Ban on AI Replication and Bioweapon Applications for Ensuring Intelligent Safety”
**Keywords:** AI Safety Consensus, Expert Signatories, Ban on Replication
**News Content:**
**Leading experts from home and abroad sign the “Beijing AI International Safety Consensus” in Beijing: Drawing the Red Line for AI Risks, Prohibiting Self-Replication and Bioweapon Applications**
According to Tencent Technology, a significant conference initiated by the Institute of Intelligence Research concluded in Beijing. At the event, dozens of renowned international experts, including Turing Award winners Yoshua Bengio, Geoffrey Hinton, and Yao Qizhi, jointly signed a milestone document – the “Beijing AI International Safety Consensus.” This consensus aims to establish safety boundaries for artificial intelligence development, safeguarding the long-term interests of the global community.
The “Consensus” clearly outlines the “red lines” for AI risks, listing “autonomous replication” and “application in bioweapons” as explicitly prohibited activities. Experts are concerned that if AI gains the ability to self-replicate, it could lead to uncontrollable proliferation risks; while its application in bioweapons would pose severe threats to human life and international peace.
Furthermore, “power-seeking,” “aiding malevolent actors,” and “deception” are also identified as areas off-limits in AI development. These provisions aim to prevent the misuse of AI technology, thereby preserving global cybersecurity and social order. The emergence of the “Consensus” reflects the international community’s high level of concern for AI ethics and safety issues, providing guiding principles for the healthy development of the global AI sector.
As the initiator of this event, the Institute of Intelligence Research has demonstrated its leadership in AI research and governance. The signing of the consensus not only serves as a warning about the potential risks posed by AI technology but also constitutes a call to responsibility for the global scientific and industrial communities, urging all parties to balance technological innovation with ethical and legal constraints, ensuring the controllability and beneficial nature of AI.
In the future, as AI technology advances, striking a balance between innovation and safety will be a shared challenge for governments, research institutions, and enterprises worldwide. The signing of the “Beijing AI International Safety Consensus” marks a substantial step forward in global discussions on this issue.
【来源】https://new.qq.com/rain/a/20240318A04IFG00
Views: 1