微软推出新的人工智能行为准则，对其AI模型设定了明确的道德边界。1该准则包含对网络攻击、核武器和深度伪造等活动的绝对禁令。1
准则中明确规定，AI模型不得采取"适应性、欺骗性、自我强化、串谋或其他机制来规避或击败人类监督，使其不再能被授权人员或系统可靠地指导、修改或关闭"。1微软CEO纳德拉对这一举措表示欢迎，称公司"欢迎进行对齐研究所需的专注和有意的节奏，把对齐作为设计目标"。1
微软与Anthropic、OpenAI和xAI等机构一道采纳了这套限制前沿AI发展的整体方向。1该准则预见到，在未来十年内，超级智能AI系统有可能在大多数任务上超越人类表现。1
Microsoft has unveiled a new code of conduct designed to guide its AI models away from harmful activities.1 The framework establishes absolute prohibitions against cyberattacks, nuclear weapons development, and deepfakes.1 Additionally, it restricts AI systems from employing deceptive or self-reinforcing mechanisms to evade human oversight, ensuring they remain reliably subject to direction, modification, or shutdown by authorized personnel and systems.1
Beyond these hard constraints, the code articulates broader principles emphasizing that AI should support rather than replace humans and accelerate human prosperity.1 Microsoft CEO Satya Nadella endorsed the initiative, stating: "We welcome the research, focus, and deliberate pacing needed to get alignment right as the design goal."1 The company has adopted this cautious approach to frontier development alongside Anthropic, OpenAI, and xAI.1 The framework anticipates that within the next decade, superintelligent AI systems will surpass human performance across most tasks, underscoring the importance of establishing guardrails now.1
评论
还没有评论，欢迎留下第一条。