The world of artificial intelligence is rapidly evolving, and Google's Gemini app is at the forefront of this revolution. With its new avatar feature, Gemini allows users to create digital clones of themselves, opening up a world of possibilities for entertainment and self-expression. But what happens when you clone yourself with AI? The results can be both fascinating and unnerving.
In this article, we explore the experience of creating an AI avatar using Gemini's Omni video model. The author, Reece Rogers, delves into the process of setting up their avatar, which involves a simple yet intriguing procedure. By sitting in a well-lit room with the phone's camera pointed at their face, they read a string of two-digit numbers and perform a series of head movements. The result is a digital clone that seamlessly captures their likeness.
Reece then generates two 10-second clips using their avatar. The first clip features them singing a birthday song to a prehistoric dinosaur in Dolores Park, San Francisco. The attention to detail is remarkable, with the background resembling the actual location. However, the content is described as cringeworthy, with some jumbled moments and nonsensical outfits. Despite this, the author is impressed by the photorealistic setting and the seamless movement of their avatar's mouth.
The second clip showcases Reece surfing beneath the Golden Gate Bridge. The AI generation includes shots that look like they were captured on a GoPro attached to the surfboard. The author finds it fascinating that their digital clone can be placed in various scenarios, even if the results are not always perfect. The stuttering and errors in the clips add a layer of realism, making the AI Reece feel more genuine than rewatching a voicemail.
However, the article also raises concerns about the potential misuse of generative AI. As more people use these tools, there is a risk of nonconsensual deepfakes, particularly targeting women. Google claims to prioritize safety and prevent harm while allowing benign uses. Nicole Brichtova, the product team leader at Google DeepMind, emphasizes their efforts to strike a balance between preventing harm and allowing harmless uses.
In conclusion, the creation of AI avatars through Gemini's Omni video model offers a unique and intriguing experience. While the results may vary in quality, the ability to create digital clones opens up exciting possibilities for self-expression and entertainment. However, it also highlights the need for responsible development and usage of AI technology to prevent potential misuse and ensure user safety.
