Large Language Models (LLMs) can be traced back to days when AIs began to understand the context of word meanings.
But the pace didn't really pick up until deep leaning and the transformers breakthrough. And the development further escalated when OpenAI introduced ChatGPT. Since then, tech companies wage an arms race, where they compete to create the best and most powerful LLMs possible.
With content creation, coding, customer service, tutoring, and many others under people's fingertips, LLM-powered AIs literally enable anyone harness superpower abilities via natural language.
Google has stepped into the battle with its own solution, Gemini, leveraging the full breadth of its technology and resources in a bold move to challenge OpenAI and some others.
And this time, Google is making yet another upgrade to its Gemini.
The company is giving its Gemini AI a new native editing tools that let users seamlessly modify both uploaded photos and AI-generated images.
With the feature, users can easily can change backgrounds, add or swap objects, or even create entirely new images—like seeing how they'd look with a different hair color.
Starting today, we're testing new image editing features in Gemini like changing backgrounds, swapping objects, and more. They’re rolling out gradually to all users. We're continuing to improve image generation, so stay tuned. pic.twitter.com/JsurD7LRck
— Google Gemini App (@GeminiApp) April 30, 2025
In a blog post, Google said that:
"Now, you'll be able to easily edit both AI-generated images and personal photos uploaded from your phone or computer. You can modify backgrounds, swap out objects, add new elements, and more. For instance, you could upload a photo of yourself and ask Gemini to show how you'd look with different hair colors."
These easy-to-use, step-by-step editing features make interactions with Gemini more dynamic and engaging—blending text and visuals effortlessly.
For example, users can ask Gemini to make up a new bedtime story about dragons, complete with custom illustrations that bring the tale to life.
The feature works for both users own images, as well as AI-generated images.
And to ensure a smooth flowing engagement, Google added that any generated and modified images should appear right within the conversation window.
According to Google, this native image editing is rolling out gradually and will be available in "over 45 languages and most countries in the coming weeks."
Great question—both! You can ask Gemini to edit AI-generated images and your own images, and it will modify the images within your conversation.
— Google Gemini App (@GeminiApp) May 1, 2025
This kind of feature is a next-level superpower, since it essentially gives anyone with the ability to create deepfakes with insane easiness.
And this is reasonable, knowing how far LLMs have advanced.
Google that knows the risk, said that images generated and edited with Gemini shall all have an invisible watermark.
Powered by SynthID from Google DeepMind, it's an advanced watermarking technology designed to embed imperceptible digital markers directly into AI-generated content—be it images, audio, video, or text. These watermarks are invisible to the human eye and inaudible to the ear but can be detected by specialized software, allowing for the identification of AI-generated material.
This technology is meant to facilitating content verification, as well as enhancing transparency, allowing users to make informed decisions about the media they consume and share.