TikTok’s latest update allows users to transform their own voice into a simulated version using artificial intelligence (AI), allowing them to use their digital voice instead of generic template speakers in their content.
The China-based social media platform is pioneering text-to-speech technology and offers users AI-generated voiceover features with preset voices. While Instagram and other social networks have also adopted this feature, TikTok goes a step further and offers custom AI voiceovers using users’ voices.
Social media expert Jonah Manzano discovered the update and posted in threads about the new “Create your own AI voice” option that some users found in TikTok’s voiceover features. Manzano also tested the tool in a separate YouTube video.
Users are asked to record themselves reading a few lines to create their own AI voice. This audio recording is then used to create a synthetic voice that mimics their voice in just 10 seconds. TikTok allows multiple attempts until the user is happy with their digital voice, which they can then apply to their clips using text-to-speech.
In addition, the new update goes a step further and includes an additional feature that translates custom voiceovers into other languages. According to Manzano’s video, the available languages are English, Spanish, Chinese, Portuguese, Japanese, and Indonesian.
TikTok was able to integrate this tool into its app thanks to its parent company ByteDance. In January, ByteDance reportedly succeeded in developing an AI system called “StreamVoice” that can imitate human voices.
SteamVoice can generate credibly accurate human voices in real time based on just a few utterances in short audio inputs. This breakthrough has made it possible to reproduce any person’s voice.
TikTok has acknowledged that its new tool, which uses lifelike voices, could be used to produce more convincing hoaxes, deepfakes and fake news, saying: “Your AI voice is private and you can delete it at any time.”
