{"id":6927,"date":"2024-04-02T09:15:05","date_gmt":"2024-04-02T01:15:05","guid":{"rendered":"https:\/\/www.1ai.net\/?p=6927"},"modified":"2024-04-02T09:15:05","modified_gmt":"2024-04-02T01:15:05","slug":"%e9%81%8f%e5%88%b6-ai-%e5%a4%b1%e6%8e%a7%ef%bc%8c%e5%be%ae%e8%bd%af%e5%8f%91%e5%b8%83%e7%b3%bb%e5%88%97%e5%b7%a5%e5%85%b7%e5%87%8f%e5%b0%91-copilot%e5%b9%bb%e8%a7%89%e6%83%85","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/6927.html","title":{"rendered":"To curb AI out of control, Microsoft released a series of tools to reduce Copilot "hallucination" situations"},"content":{"rendered":"

Behind the popularity of generative AI, security, privacy and reliability issues are becoming increasingly prominent.Microsoft<\/a>In order to prevent the occurrence of Supremacy AGI (AI that claims to control the human world),A series of solutions have been launched recently to prevent generative AI from getting out of control.<\/strong><\/p>\n

\"f5aadccb-c9cb-478e-9cb7-1b5cc203e0cc\"<\/p>\n

Microsoft said in an official announcement: "How generative AI can effectively prevent prompt word injection attacks has become a major challenge. In this attack, malicious actors try to manipulate artificial intelligence systems to do something beyond their intended purpose, such as creating harmful content or leaking confidential data."<\/p>\n

Microsoft first restricted Copilot<\/a> the number of characters to mitigate serious hallucinations. In addition, Microsoft introduced the \u201cBasic Test\u201d function, designed to help users recognize text-based hallucinations\u3002<\/p>\n

This feature will automatically detect \u201cungrounded material\u201d in text to support the quality of LLM output, ultimately improving quality and confidence.<\/p>\n

The relevant tools are introduced as follows:<\/p>\n