Google unveils AI-powered mouse pointer with Gemini integration
Google DeepMind researchers Adrien Baranes and Rob Marchant are developing an AI-enabled mouse pointer that understands context and intent, powered by the Gemini model. The pointer aims to replace text-heavy prompts with intuitive pointing and speech, allowing users to interact with pixels as actionable entities. Four design principles guide the project: maintaining user flow across apps, embracing natural language like "this" and "that", turning pixels into structured data, and adapting technology to human behavior. Experimental demos show users editing images or finding places in Google AI Studio by pointing and speaking. The technology is being integrated into Chrome and the new Googlebook laptop, with features like Magic Pointer rolling out soon. Google Labs' Disco will also test future concepts.
Key facts
- Developed by Adrien Baranes and Rob Marchant at Google DeepMind
- Powered by Gemini AI model
- Four interaction principles: maintain flow, embrace 'this' and 'that', turn pixels into actionable entities, adapt to human behavior
- Experimental demos available in Google AI Studio
- Integration into Chrome and Googlebook laptop
- Magic Pointer feature coming to Googlebook
- Google Labs' Disco to test future concepts
- Aims to eliminate need for detailed text prompts
Entities
Artists
- Adrien Baranes
- Rob Marchant
Institutions
- Google DeepMind
- Google AI Studio
- Google Labs
- Chrome
- Googlebook
- Disco