Most of us have likely experienced artificial intelligence (AI) voices through personal assistants like Siri or Alexa, with their flat intonation and mechanical delivery giving us the impression that we could easily distinguish between an AI-generated voice and a real person. But scientists now say the average listener can no longer tell the difference between real people and "deepfake" voices.

In a new study published Sept. 24 in the journal PLoS One, researchers showed that when people listen to human voices — alongside AI-generated versions of the same voices — they cannot accurately identify which are real and which are fake.

"AI-generated voices are all around us now. We’ve all spoken to Alexa or Siri, or had our calls taken by automated customer service systems," said lead author of the study Nadine Lavan, senior lecturer in psychology at Queen Mary University of London, in a statement. "Those things don’t quite sound like real human voices, but it was only a matter of time until AI technology began to produce naturalistic, human-sounding speech."

The study suggested that, while generic voices created from scratch were not deemed to be realistic, voice clones trained on the voices of real people — deepfake audio — were found to be just as believable as their real-life counterparts.