Scroll down for the explanation.
If you guessed it was fake, you’re correct! This clip was generated using AI methods. The green boxes you see on the sound wave highlight consonant bursts, which is one of the linguistic features that we found can help listeners tell if a voice is real or fake. In this clip, the prominent bursts at the end of “not” and “hot”, in combination with the overall audio quality distortion, point to it being fake.