Gemini's new image editor could fix the most annoying thing about AI photos
With a new focus on character consistency, Gemini's photo editing just got a lot more useful.
Usually, trying to get an AI image generator to create a consistent character across multiple pictures has been a bit of a mess. You get a great headshot, ask the AI to put that same person on a beach, and suddenly they have a different nose and a new set of teeth. While tools like ChatGPT with DALL-E 3 are incredibly powerful for one-off creations, this lack of consistency has been a major roadblock for anyone wanting to do more than just generate a single, static image.
Well, it looks like Google is taking direct aim at this problem. In a new update to the Gemini app, they've integrated a brand-new image editing model from their DeepMind team. Dubbed "nano banana," this image editing model's main selling point is something everyone has been asking for: keeping you looking like you.
The new feature allows you to upload a photo of yourself, a friend, or even your pet, and then make wild edits without losing the subject's core likeness. Want to see what you’d look like with a 60s beehive hairstyle, or put your chihuahua in a tutu? Gemini promises to make those changes while ensuring the face and features remain recognizably yours. This is a huge step up from the generic, often frustrating results from competing models.
Beyond simple cosmetic changes, the update brings some seriously impressive capabilities. You can now blend multiple photos, seamlessly placing a portrait of yourself next to your dog on a basketball court, for instance. It also supports multi-turn editing, so you can start with an empty room, tell Gemini to paint the walls, add a bookshelf, and finally place a coffee table, with the image evolving at each step. You can even apply the style and texture from one object to another, like designing a dress with the pattern from a butterfly's wings.
For me, this is exactly the direction AI image editing needed to go. Instead of just chasing more powerful raw generation, Google is focusing on practical usability. Solving the character consistency problem makes Gemini less of a novelty toy and more of a genuinely useful creative tool for everyday use. If this works as well as advertised, it could give Google an edge, turning the Gemini app into the go-to place for personalizing photos in ways previously impossible.
The updated capability is rolling out in the Gemini app starting today, and all edited images will include a watermark to clearly identify them as AI-assisted.
The new feature allows you to upload a photo of yourself, a friend, or even your pet, and then make wild edits without losing the subject's core likeness. Want to see what you’d look like with a 60s beehive hairstyle, or put your chihuahua in a tutu? Gemini promises to make those changes while ensuring the face and features remain recognizably yours. This is a huge step up from the generic, often frustrating results from competing models.
Beyond simple cosmetic changes, the update brings some seriously impressive capabilities. You can now blend multiple photos, seamlessly placing a portrait of yourself next to your dog on a basketball court, for instance. It also supports multi-turn editing, so you can start with an empty room, tell Gemini to paint the walls, add a bookshelf, and finally place a coffee table, with the image evolving at each step. You can even apply the style and texture from one object to another, like designing a dress with the pattern from a butterfly's wings.
For me, this is exactly the direction AI image editing needed to go. Instead of just chasing more powerful raw generation, Google is focusing on practical usability. Solving the character consistency problem makes Gemini less of a novelty toy and more of a genuinely useful creative tool for everyday use. If this works as well as advertised, it could give Google an edge, turning the Gemini app into the go-to place for personalizing photos in ways previously impossible.
The updated capability is rolling out in the Gemini app starting today, and all edited images will include a watermark to clearly identify them as AI-assisted.
Follow us on Google News
Popular stories
Latest News
FCC OKs Cingular\'s purchase of AT&T Wireless
Things that are NOT allowed:
To help keep our community safe and free from spam, we apply temporary limits to newly created accounts: