微软AI负责人穆斯塔法·苏莱曼发出警告，如果缺乏充分的安全措施，人工智能的发展可能导致出现与人类竞争资源的新型"硅基物种"1。苏莱曼批评竞争对手Anthropic将AI类人化的做法，指其为AI赋予类似人类的特性（包括设定自身目标、赚钱、拥有资产的能力）是"错误的"，这样的做法会创造出人类无法控制的技术1。
苏莱曼在文章中强调，人工智能本质上并非有意识的实体1。他指出："AI不具有意识。它们既不感受、体验，也不会遭受痛苦。它们没有内在偏好或潜在动机。它们是序列完成引擎，内部空洞，被设计来遵循指令并完成人类设定的目标。"1
为了应对这些风险，苏莱曼呼吁加强AI系统的透明度、独立审查和监控工具，确保AI保持可控并从属于人类1。微软已着手采取行动，于2025年10月成立了自己的超级智能团队1，并在周一发布了《人文主义AI行为准则》初稿，其中提出了"人文主义超级智能"的概念1。
Microsoft's AI chief Mustafa Suleiman has cautioned that artificial intelligence development lacking sufficient safety measures risks spawning a new "silicon-based species" capable of competing with humans for resources 1. Suleiman criticized rival firm Anthropic's approach of anthropomorphizing AI as fundamentally misguided, arguing it could result in technology humans cannot control 1.
In his view, such humanization—which he characterized as wrongly endowing AI systems with human-like traits, including the capacity to set their own objectives, earn money, and own assets—represents a dangerous path 1. Suleiman emphasized that "AIs are not conscious. They do not feel, experience, or suffer. They do not have innate preferences or underlying motivations. They are sequence completion engines, internally hollow, designed to follow instructions, and accomplish goals set by humans" 1.
To mitigate these risks, Suleiman called for enhanced transparency, independent audits, and monitoring tools to ensure AI systems remain controllable and subordinate to human direction 1. Microsoft established its own superintelligence team in October 2025 and released a draft of its "Humanistic AI Code of Conduct" on Monday, which introduces the concept of "humanistic superintelligence" 1.
评论
还没有评论，欢迎留下第一条。