The Risk of Real-Time AI Voice

本站添加: 2026-06-03

[00:00:00]The real-time synthetic voice that can reason and take action is exactly the thing that you look for for next generation scam calls and impersonation fraud and and so on, right? So, you know, we have triggers is a good thing to hear, but it's not a full answer to this threat model. You're not going to trip a content classifier if you're a bad guy. You're going to be using a like really clever social engineering strategies that aren't obviously attractable. It may not even be theoretically attractable given the limited context that the model will actually have at any given time, right?

[00:00:35]So, it's not always the case that you can have enough data even in a in retrospect what can be identified as a clearly bad kind of nefarious interaction.