In a significant update to its Gemini AI image generator, American tech giant Google has introduced an innovative editing tool that allows users to fine-tune specific parts of generated images. This feature, now included in the latest Android 15 Beta version, aims to provide a more customisable approach to image generation by enabling users to edit certain portions of an image with precision. However, while the tool offers exciting new possibilities, its current state in beta suggests it may not yet be flawless.

According to reports from Android Authority, Google has enhanced its AI image generation capabilities by allowing users to interact more directly with the content. Previously, users could only generate images through text prompts. The newly added feature builds on this by enabling users to highlight areas of an image and request changes.

For instance, during a demo, a user generated an image of a dog wearing a hat and sunglasses, and then requested to “change the hat with a birthday hat”. The AI swiftly adapted the image to reflect the request.

The tool’s most notable advancement is its ability to let users pinpoint exact sections of an image for modification. This offers more granular control over the final result, making it possible to customise the output in a way that was previously difficult. Despite the improvements, the feature is not without its limitations.

Moreover, the demonstration video indicated that the AI sometimes struggled to understand requests accurately, with certain edits showing inconsistencies. The process of making changes also took a noticeable amount of time, and further refinement of the tool is expected before its official launch.

Google’s move comes as part of its ongoing efforts to perfect its Gemini AI image generator, which has been in development since the platform’s launch. The feature remains in beta, and Google has indicated that users can expect improvements in both speed and accuracy as the feature matures into a full release.