Voice synthesis technology has advanced to the point where convincing synthetic audio of any individual can be generated from a few minutes of recorded speech. This technology is being used to create non-consensual synthetic intimate audio content, fake explicit phone calls, audio messages, and voice-over videos. Discovering your exposure related to intimate visual depictions can include video content with deepfake audio, and the audio itself may be relevant under existing statutes. Platform policies increasingly treat synthetic intimate audio as NCII.

How voice deepfakes are created

Voice synthesis tools including ElevenLabs, Tortoise TTS, and similar platforms allow realistic voice cloning from audio samples. Perpetrators use publicly available audio, interviews, podcasts, social media videos, to train voice models. The resulting synthetic voice can say anything the perpetrator scripts, including explicit content. Audio deepfakes are often combined with video deepfakes or used in social engineering attacks.

Understanding coverage for voice deepfakes

Discovering where your face appears in intimate visual depictions can include video content that features synthesized audio of the victim. Audio-only synthetic intimate content may relate to wiretapping statutes, fraud provisions, or harassment laws depending on how it is used. Several states have enacted specific AI voice cloning restrictions. An attorney can advise on the specific applicable statutes.

Discovery approach for audio deepfakes

For voice deepfake content embedded in video, ScanErase's biometric scan identifies the video based on the visual component. For audio-only deepfakes, ScanErase's technology finds and reports the audio content specifically. Platform content policies increasingly treat synthetic intimate audio as NCII.

Frequently asked questions

Can someone be prosecuted for creating a synthetic audio deepfake of me saying explicit things?

Potentially yes, depending on how the content is used. Using synthetic audio to defraud, harass, or defame may violate existing laws. Some states specifically criminalize AI voice cloning for deceptive purposes.

How do I prove that an audio recording is a deepfake and not authentic?

Voice forensic experts and audio analysis tools can identify synthetic audio artifacts. For your exposure report, proving the audio is synthetic is not required to assert your rights, if the content is intimate and non-consensual, it is covered regardless of authenticity.

Related deepfake resources