Google's creepy new Gemini Live Avatars want to try and make online support bots feel more human (www.techradar.com)

Google has unveiled its latest AI model, Gemini 3.8 Live, which introduces customizable avatars designed to enhance online interactions by providing a more human-like presence. Through advanced real-time speech capabilities and low-latency video generation, these avatars can lip-sync to the AI's voice, making customer service, training, and similar applications feel more engaging. Businesses can choose from a library of preset avatars or create custom options that align with their branding, such as featuring logos on uniforms. This development is significant for the AI/ML community as it showcases the intersection of AI with user experience, aiming to bridge the gap between human interaction and machine assistance. Gemini 3.8 Live also boasts impressive multilingual capabilities, supporting 97 languages with automatic switching for seamless conversations. Furthermore, a more advanced variant, Gemini 3.8 Live Extended Thinking, has demonstrated superior performance compared to competitors like GPT-Live-1 Astra while being more cost-effective for audio processing. To address safety concerns regarding AI-generated content, Google implemented SynthID watermarking in its videos, ensuring that the output is clearly identified as AI-generated, which is crucial for combating misinformation.