Summary
- The latest version of the TikTok app for Android contains strings detailing a brand-new text-to-speech feature based on AI.
- When available widely, TikTok users will be able to record their voices by reading out a piece of onscreen text.
- Additional details on privacy and safety concerns are awaited, but people will have the choice to remove their recorded voices from the TikTok Voice Library.
AI has taken the tech world by storm over the past 18 months or so, thanks largely to the emergence of OpenAI’s ChatGPT. Today, you will hardly find an app or software that doesn’t have some sort of AI capability built-in, including popular social media apps like Instagram and WhatsApp. One of the world’s most widely used content platforms, TikTok, already offers some AI features for its users. The app is now working on a new AI-based text-to-speech feature, according to a prominent source.
How to use TikTok and other data-sucking apps without giving up your privacy
Protect your privacy on TikTok and other apps with our guide
Writing for TheSpAndroid, code hunter AssembleDebug dug into the latest version of the TikTok app for Android to find code referencing something known as the “TikTok Voice Library,” plus a new option to “Create your voice with AI.” There’s no formal name for this feature just yet, per the source, but these terms pretty much explain what they’re supposed to do.
AssembleDebug even managed to get some of the features to work, giving us a glimpse of the introduction screen, which explains the process: Create an AI version of your voice in just 10 seconds! You’ll be able to use it with text-to-speech in TikTok videos.
A handy feature, but with some risks involved
Tapping Continue on the intro screen takes users to the first of four TikTok Voice Library Terms pages, explaining how the voice recordings are managed by the platform while also detailing how a recorded voice can be removed from TikTok’s servers. Hitting Agree takes users to the recording page, which appears blank for now, possibly due to the feature being in its early stages of development.
But when available, users will be asked to read out a piece of onscreen text, which will help TikTok recognize the voice and eventually make it available in the TikTok Voice Library for others to use. AssembleDebug speculates that users will likely have to record their voices just once, though there could also be an option to rerecord the voice.
There will always be concerns about the potential misuse of an AI-based text-to-speech tool such as this one, particularly at a time when misinformation runs rampant on online platforms. We hope TikTok will provide some additional details on how it will combat these menaces. However, since this text-to-speech tool is still under development, we don’t expect the company to come out with a word until the feature is ready for a broader release.
