Insta360 will launch an AI voice assistant feature for the GO Ultra thumb camera on August 7th. The feature uses different large model solutions based on region: the Alibaba Qwen large model is integrated in mainland China, while Google Gemini is used in Hong Kong, Macau, Taiwan, and overseas regions. The translation results can be directly announced through the device's speaker.

According to the introduction, this AI assistant is based on Insta360's self-developed technology, integrating the multimodal capabilities of Qwen large model and the photo Q&A function of the Qwen App, using a terminal-cloud collaborative architecture. The terminal side is responsible for tasks such as voiceprint recognition and intent judgment, while the cloud completes complex processing such as intelligent Q&A, mode switching, and translation.

In terms of features, users can activate the assistant by saying "Hey Kira" or long-pressing the capture button while the device is off. It supports face-to-face translation, photo-based Q&A, and voice-based Q&A. Among these, face-to-face translation can be used in scenarios such as asking for directions, ordering food, and shopping; photo-based Q&A, after obtaining authorization, can identify content such as road signs, buildings, and dishes in the lens and provide analysis.

Liu Jingkang, founder of Insta360, stated that this feature may "redefine the thumb camera." This upgrade also shows that AI large models are accelerating into the consumer hardware field, driving smart cameras from being shooting tools to AI terminals with understanding and interaction capabilities.