344
Productivity & Workflow355
Automation & Workflow224
Software Development251
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative170
Photography & Imaging156
Customer Support131
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
TikTok is testing an opt-in feature that scans for AI-made deepfakes using a creator’s face or voice and lets creators report them.
In short: TikTok is testing an opt-in tool that looks for AI-made deepfakes that copy a creator’s face or voice.
TikTok is testing a new feature it calls an AI likeness detection tool. It is meant to help creators find videos that may be using their likeness without permission. “Likeness” here means a person’s face or voice, or both.
The test is starting with some creators in the US, according to TikTok US spokesperson Zachary Kizer. The feature was first spotted by social media consultant Matt Navarra.
Creators who join the test and want to use the tool must first verify their identity with Jumio, a company that checks identities online. TikTok says this includes a real-time selfie scan and an ID check. Kizer told The Verge that TikTok does not keep the ID documents, and that facial information is only used to match a creator’s likeness and help spot possible unauthorized uses.
After verification, TikTok scans for AI-generated content that could be using a creator’s likeness. The creator can then review what TikTok finds and report posts or accounts they believe are unauthorized.
TikTok is not alone in working on this type of feature. YouTube has been developing a similar tool, and it has recently made its version available to all adult users.
AI deepfakes are getting easier to make, and they can be used to impersonate people. A tool like this works a bit like a plagiarism checker, but for someone’s face and voice. If it works well, it could help creators spot impersonations faster and make it simpler to report them.
Source: The Verge AI