Google’s Gemini app now offers a subscription-based feature that creates photorealistic digital avatars of users, enabling them to star in AI-generated videos that closely mimic their real-life appearance and mannerisms.
What happened
The Gemini avatar tool, powered by Google’s Omni video model, allows adult subscribers of the $20-per-month Gemini AI Pro plan to generate short video clips featuring a digital clone of themselves. Users create the avatar by recording a brief calibration video with their phone’s camera, involving face scanning and head movements. Once set up, the avatar can be inserted into videos generated by AI prompts, placing users in various scenarios.
Reece Rogers experimented with the tool and produced two 10-second clips: one where his avatar sang “Happy Birthday” to a CGI dinosaur at Dolores Park in San Francisco, and another where he appeared to surf beneath the Golden Gate Bridge, albeit wearing unexpected denim rather than a wetsuit. The backgrounds were photorealistic and location-accurate, showcasing Google’s mapping technology integration.
Despite some glitches, such as stuttering, erratic props, and mismatched outfits, the avatar’s facial features and movements were strikingly lifelike. The experience prompted mixed feelings of fascination and unease, underscoring the uncanny quality of digitally cloned personas.
Why it matters
This development illustrates the rapid advancement of generative AI in creating hyper-realistic digital replicas of individuals, expanding creative use cases in video production and personal content creation. At the same time, it raises important safety concerns, especially around consent and potential misuse of deepfakes.
Google emphasizes safety measures to prevent harm and restricts avatar creation to consenting adult users only, differentiating itself from other AI tools that previously allowed decentralized likeness generation. The integration of real-world geography in AI-generated video backgrounds also demonstrates how tech giants leverage their broader data assets to enhance AI realism.
Background
Avatar-based AI video generation is an emerging technology aiming to personalize digital content by replicating real individuals’ likenesses and behaviors. Google’s Gemini app adds to a growing field of AI tools enabling users to produce synthetic media, following earlier models like OpenAI’s now-discontinued Sora app. As generative AI becomes more capable, tech companies are balancing innovation with ethical considerations regarding privacy, consent, and potential misuse.
Sources
This article is based on reporting and publicly available information from the following source:
Read more Artificial Intelligence stories on Goka World News.