Google has expanded its “AI Mode” search feature to support image understanding. The update enables users to upload or take photos that AI Mode can analyze to provide relevant search results. According to Ryan Whitwam at Ars Technica, the feature uses a custom version of the Gemini large language model working alongside Google Lens to identify specific objects in images. This “fan-out technique” allows the system to make multiple sub-queries based on visual content. Initially limited to Google One AI Premium subscribers, the company is now extending access to “millions more Labs users in the US” who aren’t paying for AI features. Google reports that users are entering approximately twice as much text in AI Mode compared to traditional searches, suggesting people may need to provide more context when using the AI-powered system.