Google will peel back a new era of AI images with Nano Banana 2

Google's latest AI image generator carries a peculiar name that sounds more like a smoothie ingredient than cutting-edge technology. Nano Banana 2 recently surfaced in unexpected preview glimpses, revealing a fundamental shift in how artificial intelligence creates pictures.

Unlike its predecessor, this new model thinks before it draws. The system plans an image, creates it, then scrutinizes its own work for mistakes before trying again. This self-correction loop continues until the AI deems the result acceptable.

The approach mirrors human creativity more closely than previous AI tools, which simply generated images in one shot. Early previews show cleaner lines, sharper angles, and fewer telltale AI errors that typically plague machine-generated art.

Beyond technical improvements, Nano Banana 2 promises better control over angles, viewpoints, and coloring. Users can even fix text within images without affecting the rest of the composition.

Code references already hint at "Nano Banana Pro," suggesting Google plans premium versions for high-resolution tasks. The quirky naming aside, Google appears intent on transforming AI image generation from a novelty into a genuine design assistant.