Insta360, a Chinese tech company, has launched an AI voice assistant for its GO Ultra thumb camera, adopting different large model solutions based on regions: the Alibaba Qwen large model is integrated in the China mainland, while Google Gemini is used in the Hong Kong, Macau, Taiwan, and overseas markets. This feature is core to Insta360's self-developed technology, combining Qwen's multimodal capabilities and photo-based question-answering functions. On-device, it performs voiceprint recognition and intent identification, while the cloud handles complex tasks such as answering questions, switching modes, and translation. The translated results can be directly broadcast through the device's speaker. Insta360 founder Liu Jingkang stated that this will "redefine the thumb camera," and he mentioned that it might become the largest personal AI assistant in terms of inventory and sales volume, apart from Meta.

image.png

The AI voice assistant integrates three core capabilities: face-to-face translation, image-based question-answering, and voice-based question-answering. Users can activate it by saying "Hey Kira" or by long-pressing the shutter button when the device is off. Face-to-face translation supports bidirectional real-time translation, covering scenarios such as asking for directions, ordering food, and shopping; image-based question-answering captures the lens view after authorization, identifying elements like road signs, buildings, and dishes, and providing voice feedback; voice-based question-answering does not use the camera, directly responding to general information queries such as weather and routes. The combination of these three features upgrades the thumb camera from a simple photography tool into an intelligent terminal that can perceive and converse.

A New Player Enters the Wearable AI Arena

GO Ultra was released last August at a price of 2598 yuan. It features a lightweight 53-gram cube-shaped body, equipped with a 1/1.28-inch large sensor and a 5nm AI chip, supporting 4K60fps high-frame video recording. By introducing an AI voice assistant into this form factor, Insta360 has essentially opened up a new path in the wearable AI hardware market, distinct from smart glasses and smart speakers. The lightweight wearing advantage of the thumb camera naturally makes it suitable as a full-time AI assistant carrier, while the built-in camera provides a hardware foundation for multimodal interaction.

Selecting different large models based on regions reflects the practical considerations faced by domestic hardware manufacturers in their global expansion. Using Alibaba Qwen in the China mainland ensures compliance and data localization, while using Gemini in overseas markets provides broader general capabilities. This strategy may also serve as a reference model for other hardware companies expanding abroad. As Meta's smart glasses, OpenAI's ice ball devices, and Rabbit's handheld devices enter the market, the competitive landscape for personal AI assistant hardware is accelerating. Whether Insta360 can secure a place in this arena depends on continuous optimization of the AI experience and deep development of the ecosystem.