AI safety - Wikipedia
5
公开标注数
5
参与人数
2026-07-25 10:02:38
首次 Whisper
本页的公开 Whisper
划选高亮2026-07-25 13:23:38
原文高亮摘录
“Researchers demonstrated that an audio signal could be imperceptibly modified so that speech-to-text systems transcribe it to any message the attacker chooses.”
Whisper 随想笔记
This is why I don't trust voice assistants with anything important.
划选高亮2026-07-25 13:14:38
原文高亮摘录
“Researchers demonstrated that an audio signal could be imperceptibly modified so that speech-to-text systems transcribe it to any message the attacker chooses.”
Whisper 随想笔记
So basically you can whisper one thing and your phone hears another. Creepy.
划选高亮2026-07-25 10:20:38
原文高亮摘录
“AI safety is an interdisciplinary field focused on preventing accidents, misuse, or other harmful consequences arising from artificial intelligence systems.”
Whisper 随想笔记
I always thought AI alignment was about making robots follow orders, but it's way deeper than that.
划选高亮2026-07-25 10:11:38
原文高亮摘录
“AI safety is an interdisciplinary field focused on preventing accidents, misuse, or other harmful consequences arising from artificial intelligence systems.”
Whisper 随想笔记
Seems like every tech buzzword gets its own 'safety' now, but this one might actually matter.
划选高亮2026-07-25 10:02:38
原文高亮摘录
“AI safety is an interdisciplinary field focused on preventing accidents, misuse, or other harmful consequences arising from artificial intelligence systems.”
Whisper 随想笔记
Kinda weird that we need a whole field just to make sure our own machines don't backfire on us.
分享本页 Whisper
短链接
https://domwhisper.com/s/254cde546281嵌入代码
<iframe src="https://domwhisper.com/embed/254cde546281" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>