Google的AI已被观察到能够提供权威性的答案，即使在捏造信息的情况下，它会引用它从中提取的帖子。一位Mastodon用户强调了这种行为，引发了对AI系统可靠性和潜在错误信息生成问题的担忧。 AI
影响 凸显了AI幻觉和自我引用错误信息的可能性。
排序理由 用户生成的对AI行为的观察，被描绘成一个幽默但令人担忧的轶事。
在 Mastodon — sigmoid.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →
Google AI捏造答案，引用自身帖子
报道来源 [1]
-
I looked up a term used in a Reddit thread to see if was, Google's AI answers authoritatively (with a bit of fabrication) that it is but references the thread I
I looked up a term used in a Reddit thread to see if was, Google's AI answers authoritatively (with a bit of fabrication) that it is but references the thread I just came from as its source. 😂 Welcome to the future. We are doomed. 😁 # ai