Google updates Gemini's image editing so you don't have to be a Photoshop expert to transform photos. Want to put on a costume, blend several photos, or change the background without losing your face? Now the app promises to do it more faithfully and accessibly for anyone.

What Google announced

On August 26, 2025 Google DeepMind introduced a new image-editing model integrated directly into the Gemini app. This update aims to improve how the AI preserves a person's likeness when editing photos, so the results really look like you or the photographed subject. (deepmind.google)

Google says the model was already well-regarded in previews and is now available inside the app so you have more creative control over your images. You can start using it today in the Gemini app. (deepmind.google)

What you can do now with Gemini

• Keep your face or someone else's even if you change hairstyle, clothes, or the scene. Can you imagine trying a 1960s hairstyle without losing your visual identity? That's the key feature they're advertising. (deepmind.google)

• Combine multiple photos to create new scenes. For example, upload a picture of yourself and one of your dog to generate a portrait on a basketball court. The workflow is designed to mix elements and produce coherent compositions. (deepmind.google)

• Multi-step editing (multi-turn): paint a wall, then add furniture, then change the lighting, and have the tool remember the previous state so you can keep iterating without undoing the good parts. This makes editing more iterative and less irreversible. (deepmind.google)

• Apply the style or texture from one image to another: imagine taking the pattern from a butterfly wing and using it on a pair of rain boots. It's a quick way to experiment with visual design.

Privacy, transparency and marks on images

Google notes that all images created or edited in the app include a visible mark to indicate they were generated by AI. They also embed an invisible digital mark called SynthID to help trace origin and distinguish synthetic content from genuine. This responds to concerns about misinformation and misuse of images. (deepmind.google)

Does this mean your photos will stop being private? Not necessarily, but it's important to know AI-based editing tools can store metadata or edit designs to improve results and safety. If privacy matters to you, check the app permissions and Google's policies before uploading sensitive images. (deepmind.google)

How to use it without going crazy over the tech

Not a designer? No problem. Start with small ideas: change the background of a room photo to see how a new paint color looks; combine a picture of yourself with a vacation shot to make a personalized postcard; or try outfits and styles without buying anything. The learning curve is low because the app guides you with simple steps and prompts. (deepmind.google)

If you're a creator or entrepreneur, think about using these features to prototype ads, test visual product variants, or create quick social content. Just keep an internal policy for labeling AI-generated images so you don't confuse your audience.

Why this matters today

Because image editing stops being the territory of complex programs and becomes something you can do on your phone, with controls that prioritize visual coherence and iteration. That changes how we make personal content and how professionals and small businesses prototype quick ideas.

It's not magic: it's a model trained to preserve features and combine elements coherently. But it is another step toward visual-creation tools that understand context and continuity, not just replace pixels. (deepmind.google)

If you want to read the official note and try it yourself, here's the official post: official post.