Thanks. What we've shared here is a demo tool to show our new speech model that can clone a voice with few seconds of audio. You can try that with English or non-English recordings, but the generated voice can only speak English at the moment. If you are looking for high-fidelity cloning, you can sign up and try it in our app here - https://play.ht/voice-cloning/
High-fidelity cloning requires at least 20 mins of good quality audio. The more the better.
We launched the playground.play.ht in beta to share the new speech model we are working on. We've been operating play.ht for a while and have teams from these companies using the platform.
The intention for this playground was to let people try the model. We actually have auto moderation on the user facing platform (https://play.ht/) and malicious text gets blocked and the user get flagged.
You are right, the technology will become ubiquitous, therefore, at least for platforms like us, it's a responsibility to have countermeasures and safeguards to prevent abuse and harm people. There'll always be people who will find ways to abuse but making it more and more difficult and evolving on that seems like a way forward.
We have these measures in place and are working on others to make sure the technology is used towards the betterment of humanity.
1/ Auto moderation on text to block harmful/malicious speech.
2/ As someone pointed out in the comments, we had a manual review process in place where the user is required to read out a consent and a member from Play.ht would review it before approving the voice. We're working on improving and adding this back.
3/ The user facing service is paywalled so we don't allow everyone in.
4/ Users trying to create malicious content are flagged and reviewed.
5/ A classifier to detect AI generated speech
You are right, and unfortunately that is a possibility, and we are working on having measure in place to guard against such attempts. We have auto moderation on the input text that will block such audio being generated. Such users are flagged in the system.
Couldn't agree more with your comment. We are working on counter measures like manual verification of voice, a classifier to detect cloned speech, etc. As of now we have auto moderation in place that detects and blocks hate/harmful speech.
Thanks for the feedback. We certainly want to support gaming studios of all sizes and are working with few of them to understand the workflows. What we've seen is not everyone wants a high-fidelity clone (which cost more), most of their voiceovers can be done with zeroshot cloning (quick clones that don't cost much).
The latency is not real-time yet but we're working on getting it to near real time. Regarding controlling the voice, we've added a few params like rate, voice guidance and temperature but for the most part the emotion is dependent on the text for now.
High-fidelity cloning requires at least 20 mins of good quality audio. The more the better.