Link: Zoom plans to launch a feature in 2025 that lets users create an AI-animated, photorealistic avatar with a head and arms from a video they record of themselves (Kyle Wiggers/TechCrunch)

Zoom is set to release AI-animated, photorealistic avatars next year, transforming user-recorded video clips into digital doubles that speak scripted lines synced with generated audio. Announced at Zoom's annual dev conference, these avatars aim to make interactions more efficient by allowing asynchronous communication.

According to Smita Hashim, Zoom's chief product officer, these avatars are designed to streamline video content creation and enhance productivity.

However, the technology also raises concerns about deepfake risks, where AI is used to convincingly replicate individuals’ appearance and voice.

Some companies like Microsoft and Tavus have developed similar technologies but implemented strict controls to prevent misuse.

Zoom has promised to incorporate "numerous safeguards" including advanced authentication and watermarking, though specifics remain vague.

Critically, the potential for misuse remains high as deepfakes proliferate, causing misinformation and scams that have already caused significant financial losses.

While regulatory efforts continue, Zoom's commitment to ensuring the integrity and security of its avatar feature will be crucial to its success when it launches as part of a premium package in 2025. #

--

Yoooo, this is a quick note on a link that made me go, WTF? Find all past links here.