**微软推出系列工具，强化Copilot安全防护以抵御生成式AI失控风险**
为应对日益增长的AI安全挑战，尤其是近期Supremacy AGI等事件引发的公众关切，科技巨头微软公司近日宣布了一系列措施，旨在确保其Copilot工具的安全性和可控性。微软此次升级主要针对防止生成式AI出现“幻觉”情况，即AI生成的内容与现实世界严重脱节，可能导致误导或错误决策。
首先，微软实施了字符数量限制，以减少Copilot在生成内容时可能出现的过度或不准确的“幻觉”现象。这一举措旨在确保AI的输出更聚焦、更精确，从而降低用户接触到不实或误导性信息的可能性。
其次，微软引入了一项名为“基础检测”（Groundedness Detection）的创新功能。该功能旨在帮助用户辨别 Copilot 生成的文本是否基于坚实的事实基础，以识别和过滤掉潜在的文本幻觉。通过这一工具，用户将能更好地评估和验证AI生成内容的真实性和可靠性。
微软的这些举措展示了其在AI伦理和安全领域的领导地位，也是对保障人工智能技术健康、负责任发展的积极响应。随着AI技术的广泛应用，确保其安全性和可控性已成为行业内外共同关注的重要议题。微软的行动无疑为其他科技公司设定了更高的标准，推动整个行业共同探索和实施更有效的安全策略。
英语如下：
**News Title:** “Microsoft Launches AI Security Tools to Counter Copilot Hallucinations and Ensure Generative AI Control”
**Keywords:** Microsoft, Copilot, AI Security
**News Content:**
In response to the growing AI security challenges, particularly公众 concerns arising from incidents like Supremacy AGI, tech giant Microsoft has recently announced a series of measures aimed at ensuring the safety and controllability of its Copilot tool. The company’s updates primarily focus on preventing “hallucinations” in generative AI, where the AI-generated content significantly deviates from reality, potentially leading to misdirection or incorrect decisions.
Firstly, Microsoft has implemented character limits to reduce the likelihood of overblown or inaccurate “hallucinations” when Copilot generates content. This move is designed to ensure AI outputs are more focused and precise, thereby decreasing the chances of users encountering false or misleading information.
Secondly, the company has introduced an innovative feature called “Groundedness Detection.” This function helps users assess whether the text generated by Copilot is grounded in solid factual basis, identifying and filtering out potential textual hallucinations. With this tool, users will be better equipped to evaluate and verify the authenticity and reliability of AI-generated content.
Microsoft’s actions demonstrate its leadership in AI ethics and security and are a proactive response to fostering the responsible development of AI technology. As AI adoption widens, ensuring its safety and controllability has become a critical issue for both the industry and society at large. Microsoft’s steps set a higher benchmark for other tech companies, encouraging the industry as a whole to explore and implement more effective safety strategies.
【来源】https://www.ithome.com/0/759/553.htm
Views: 1