Google, the technology company, has expanded the capabilities of its generative artificial intelligence model, Gemini, with new functionality. The tool can now generate interactive simulations, allowing users to visualize and practically test complex physical or scientific phenomena. This new development represents a shift from static text responses and images to dynamic content where parameters can be changed in real-time.
When asked to explain a complex topic, Gemini previously generated primarily text analyses or static diagrams. The newly integrated simulations are intended to foster a better understanding of the subject matter through user engagement. Users can actively intervene in the visualization, change settings, and observe the model's subsequent reactions. Interactive elements include control sliders, the ability to rotate the view in 3D space, and a zoom in/out function.
Practical application was demonstrated, for example, on the interaction between the Earth and the Moon. In such a case, the simulation allows users to adjust the rotation speed of the celestial bodies or enable the display of orbital paths and tidal forces. Similarly, the tool can be used in other fields outside astrophysics, making it a potentially powerful aid for education and self-study in various scientific disciplines.
However, activating these simulations is not automatic for every query. To evoke them, specific terms or commands must be used in the conversation. The tool is activated especially when using keywords such as “create simulation,” “visualize,” or “show model.” The interactive part itself does not open immediately but is offered via a “Show visualization” button that appears at the end of the generated text explanation.
The introduction of interactive models is part of Google's broader strategy to transform Gemini into a more comprehensive educational platform. Although the function is still evolving and may not be available for all types of queries, it shows the direction in which the visual interpretation of data within AI models will move in the near future.