ElevenLabs says it can detect its own voices, does that help with other tools?
podcaster here. i came across ElevenLabs' AI speech classifier, where you upload a clip and it tells you the probability it was made with their tech. i tried it on a couple of my own episodes and it said not AI, which is nice to see.
Cite
Global 100 Forum, "ElevenLabs says it can detect its own voices, does that help with other tools?", https://global100.org/forum/q/elevenlabs-says-it-can-detect-its-own-voices-does-that-help-with-other/, accessed 2026-11-14.but now i am wondering what that actually tells me. if someone cloned my voice with a different tool, or a self-hosted open model, would a classifier built by one company catch it? and if it says "not detected" on a clip, can i treat that as "real person"?
i do a lot of audio and i would like to be able to give listeners (and my mum, who gets scam calls) a simple answer to "how do i know this is really me". is a vendor detector part of that answer, or is it more of a marketing thing?
Write something first.
Give people something to work with: at least 30 words on what happened and what you tried.
That is too long. Keep it under 6,000 characters.
Write the question as the title, 15 to 140 characters, no links.
Pick a category.
Add a name (2 to 40 characters, no links).
That email address does not look right.
That was quick. Read the thread, then try again.
The form expired. Reload the page and post again.
Something went wrong with the form. Reload and try again.
Please complete the check and post again.
Limit reached for now. Try again later.
This thread is closed to new replies.
This thread no longer accepts replies.
Something went wrong with the form. Reload and try again.
Post a reply or question first (name and email), then this browser can vote, edit and accept answers.
You cannot vote on your own post.
Only the author (within 30 days) or the forum team can do that.
That email already has an account. Use your sign-in link instead.
New accounts are paused for the moment. Try again later.
That was already posted.
1 answer
- Sam WhitlockMember ·
Vendor-owned detection is the easy case: the company knows its own model and its own artefacts, and can in principle design the output to carry a mark. Detecting audio from systems you have never seen is the hard case, and that is the one your listeners care about. Different problems, and success on the first says little about the second.