Google DeepMind Reimagines the Mouse Pointer with Gemini AI

The AI Pointer Project

On May 12, 2026, Google DeepMind unveiled an experimental project to transform the standard mouse cursor into an AI-powered, context-aware tool — what the company describes as the first major rethinking of the pointer since Doug Engelbart introduced it in 1968.

How It Works

The AI-enabled pointer is powered by Gemini and continuously understands what is directly beneath the cursor: a specific word, image, code block, or UI element. Instead of copying content into a separate chatbot window, users can point at something on their screen and issue natural language or gesture commands. In a demo, pointing at an object and saying "move this here" causes Gemini to recognize the context and execute the action in place.

Try the Prototype Now

Interactive demos are available in Google AI Studio, allowing developers and users to experiment with the technology today. Demis Hassabis, CEO of Google DeepMind, called the result "pretty magical" and encouraged the community to try it. The Magic Pointer feature is also planned for the upcoming Googlebook laptop platform and for Gemini in Chrome.

Broader Implications

The AI pointer is not an incremental UX improvement — it represents a fundamental shift in human-computer interaction. If AI can act on any screen element the user points at, multi-app workflows collapse into direct, in-context actions. Further announcements are expected at Google I/O 2026 (May 19–20).