Gemini New Features: Circle, Draw Fast – AI Handles Everything

A recent leak has caused a stir in the tech community: Google is developing a brand new feature for its Gemini artificial intelligence platform that will allow users to simply circle or doodle on photos to ask AI to analyze, edit, or remove unwanted details. This is considered a breakthrough, combining Gemini's image analysis power with the intuitive and convenient "Circle to Search" feature.

1. Promising new feature just leaked

According to Android Authority, during the process of "dissecting" the Google application version 16.42.61.sa.arm64, developers discovered a new markup tool hidden in the source code. This feature allows users to draw directly on images after loading them into the Gemini interface. This is a sign that Google is preparing for a major improvement in the way people interact with AI through images.

Previously, users could upload a photo to Gemini and ask general questions like “what food is this?” or “describe this photo.” However, the limitation was that the AI ​​could only analyze the entire image, not understand which specific part the user was referring to. The newly leaked feature will change that. Instead of just entering text, users can “talk” to the AI ​​using visual actions: circling an object, underlining a line of text, or highlighting an area they want to focus on. Once the photo has been marked up, users can send requests like “delete this part,” “describe the object I just circled,” or “find information about food in this area.”

The screenshots that were discovered show that Gemini may be testing a small drawing tool that allows for color and line thickness selection. While the interface is still rudimentary, the initial test results have surprised many. Android Authority said that they tried asking Gemini to “remove the highlighted part from the photo” and the AI ​​returned “surprisingly perfect” results. The processed photo left almost no trace of editing, proving that the image processing and content generation capabilities of the Gemini 2.5 Flash Image (Nano Banana) model have reached a very sophisticated level.

If this information is confirmed, Gemini will not only be a Q&A or content creation tool, but also become a powerful image editing platform. Users can circle a dish in a photo and ask for a recipe, mark a building to ask for historical information, or simply select an unwanted object in a photo and ask the AI ​​to remove it. All of this is done intuitively, quickly and easily, without requiring advanced photo editing skills.

Although Google has not officially announced it, the source code contains code related to “drawing tools”, “highlight region” and “erase target”, leading experts to believe that it is only a matter of time before this feature appears in the upcoming Gemini update.

2. Gemini will soon be able to localize objects for AI to analyze and delete them

In Android Authority’s APK teardown, the image highlighting tool appears as an experimental feature that hasn’t been enabled yet. When users upload a photo to Gemini, a drawing interface appears that lets them circle the area of ​​the image they want the AI ​​to focus on. This “object sizing” is described in the source code as “highlighting part of an image allows you to direct Gemini’s attention to it for analysis.”

This shows that Google is aiming to help AI better understand the context of images based on visual instructions. Instead of having to write long words like “delete the man standing in the left corner of the photo”, users just need to draw a circle around the person. AI will immediately understand that the selected area is the object to be processed.

Initial tests have been promising. When a tester circled a specific area in an image and sent the command “remove circled area,” Gemini naturally recreated the background. Not only did the AI ​​remove the object accurately, but it also used scene reconstruction techniques to restore the background detail, making it look as if no edits had been made. This is a big step up from the “content-aware fill” or “object removal” tools in traditional photo editing software, which often leave behind errors or blurry edges.

This capability is supported by the Gemini 2.5 Flash Image (Nano Banana) model : an optimized variant of the Gemini model specializing in image processing. Nano Banana was designed by Google to provide faster response speed and higher efficiency on mobile devices. It is capable of understanding image context, recognizing objects, and performing image editing or reconstruction operations with very high precision.

From a user experience perspective, this feature is like an upgraded version of popular AI photo editing tools, but integrated deeply into Gemini. You can imagine that with just a circle and a few words, your photo will be processed in a snap. This makes the line between professional design and ordinary users more blurred than ever.

3. Future of convenience and impact on user design

Gemini's photo localization feature is not only a fun tool for the average user, but also a game changer for content creators, graphic designers, and media professionals.

For content creators, being able to interact directly with AI through images will open a new era. You don’t need to describe things in a complicated way, you don’t need to use complicated photo editing software. Just circle the area that needs to be adjusted, write the request and let AI do the rest. This is the perfect combination of technology and creative intuition, helping users save time while still achieving the desired results.

Imagine you’re writing a travel blog post and you have a landscape photo with a stranger in the frame. Previously, you would have to open Photoshop or other specialized editing tools to remove the extra detail. Now, just open Gemini, circle the person, write “remove person in circled area” and a few seconds later, the photo is perfect.

For designers, this feature can completely change the workflow. Instead of using multiple separate software to process images, they can complete a large part of the work right in Gemini. From removing backgrounds, changing details, to analyzing layouts, it can all be done by combining text and visual manipulation.

In particular, the ability to analyze image content by region is also a big benefit for UX/UI designers. When you want to check the contrast between parts of the interface or want AI to suggest specific edits for a small area, you just need to circle that area and ask Gemini to evaluate it. AI will analyze the selected area separately instead of the entire image, making the feedback more accurate and practical.

Beyond design, the tool is also useful for education and research. A teacher can circle a structure in a microscope image and ask Gemini about its biological characteristics. An architecture student can mark a building detail in an image and ask the AI ​​to analyze its design style. A reporter can circle a small detail in a crime scene photo and ask Gemini to identify the object or symbol that appears.

4. What is clear and what is still a guess

From the information that has been analyzed, it can be confirmed that the image markup tool has indeed appeared in the Google application source code. That means that the development team is in the internal testing phase and this feature can be activated in the near future.

Some elements like the interface, color tools, drawing gestures, and circle confirmation have been discovered. However, the official release date, availability on different devices, and the scope of the functionality are still unknown. Google has a habit of testing on limited Pixel devices before expanding to the entire Android ecosystem.

It’s also unclear whether the feature will be built directly into the standalone Gemini app or appear as an extension within “Circle to Search.” It’s also possible that Google will merge the two so users can highlight any image on their phone screen and then request analysis directly without switching apps.

Despite many predictions, experts agree that this is a reasonable step in Google's strategy. As multimodal AI models continue to develop, helping users interact with AI through intuitive actions will become an inevitable trend. And Google, with its strong image recognition platform, is fully equipped to lead this trend.

5. User experience design and instructional content

From a product design perspective, adding the ability to “draw a circle” in Gemini was a complex but interesting UX problem. The user base can come from many different groups, from professional creators to general users, so the interface had to be both intuitive and powerful.

Drawing and scribbling on images seems simple but requires sophisticated design to avoid confusion or mistakes. Users need to feel instant feedback when drawing, the scribble needs to be soft, natural, and not lag. When the AI ​​receives the command, the interface should display an effect to show that the scribble has been accurately identified.

From a content perspective, Google may also need to develop a visual tutorial system to help users understand how to use it. Short descriptions, illustrations, or demo videos will help the feature gain adoption quickly. For creative content writers, this is a great opportunity to create tutorials, articles, and videos that share tips on using Gemini in design, photo editing, learning, or communication.

In the near future, there will likely be a slew of courses, articles, or in-depth documentation on “how to use the drawing tool in Gemini,” “applying AI localization to image creation,” or “integrating Gemini into modern design workflows.” Content creators who get to grips with the tool early will have a huge advantage, as they can educate their communities and develop useful content that incorporates the new technology.

6. Conclusion

The appearance of the doodle feature in Google Gemini, although just at the level of a leak, is still considered one of the most important steps in the near future of AI technology developed by Google. It opens up a completely new interaction possibility, where people are no longer limited by words or text commands but can communicate with artificial intelligence with the most natural intuitive operations.

Gemini is getting closer to its goal of becoming a universal AI system that can understand human language, images, sounds, and actions. And if it works as expected, it will not only change the way we interact with AI, but also reshape the way we create, learn, and communicate with the digital world