"Talking Faces" Technology by Microsoft- I Put It To The Test Versus a Human

Опубликовано: 17 Июнь 2026
на канале: Deep Learning with the Wolf
1,707
38

Today we explore VASA (Visual Affective Skills Animator), a groundbreaking AI technology from Microsoft.

This technology has not been launched yet. VASA can create incredibly lifelike talking face videos from just a single photo and an audio clip, pushing the boundaries of what's possible with AI-generated avatars. Note: I did this in my makeshift office (my son's old bedroom) and used a green screen (our green Christmas tablecloth). Hey, I'm not MacGyver, but I used what I had on hand. My goal was to somewhat match the background of the Microsoft avatars.

Here is a link to Microsoft's launch page for VASA-1. https://www.microsoft.com/en-us/resea...

I do a side-by-side comparison of a Microsoft VASA avatar and a real human to showcase just how lifelike these AI-generated faces can be. (Spoiler alert: I'm the real human. Is it weird that we have to say "real" before humans now?)

#VASA #ArtificialIntelligence #MachineLearning #AvatarAI #SyntheticMedia #DeepLearning #DeepFakes #Microsoft #Research #Ethics #AIPolicymaking #Journalism #TechNews #YouTube #Podcast #DigitalLiteracy #Education #Innovation #FutureTech