Google recently made its experimental Gemini 2.0 models available to everyone for free through Google AI Studio.
This exciting development brings powerful image editing capabilities to users without design experience.
Let’s dive into what he discovered!
What Makes This Special?
Unlike other services, Google DeepMind’s “conversational editing” allows you to edit images in one simple window.
You don’t need to select specific areas or use complex tools like Photoshop- just type what you want to change, and the AI handles the rest.
How to Use It Yourself
Want to try it out? It’s simple:
- Go to Google AI Studio
- Choose the Gemini 2.0 Flash Experimental model
- Upload an image
- Write what you want to change
What Can It Do?
I tried several different editing tasks to put the system through its paces:
1. Colorizing Black and White Photos
I uploaded an old black-and-white photo and simply asked the AI to colorize it. Within just 7 seconds, I had a fully colorized version! The AI added realistic colors to the scene while maintaining the integrity of the original image.
2. Adding Objects
Next, I wanted to see how well it could add new elements. I uploaded a photo of a table and asked for flowers to be added. The results were impressive – the flowers looked natural and were placed appropriately on the table.
3. Changing Backgrounds
For my final test, I uploaded a selfie and asked to change the background to space. The result was seamless – my image was cleanly extracted and placed against a realistic space background.
Some Limitations I Noticed
The system isn’t perfect yet. Some of my attempts produced mixed results, particularly with complex images.
I noticed that sometimes the AI seems to repaint portions of photos rather than truly editing them.
Also, since this is still an experimental launch, edited images currently have lower resolution. Google will likely improve this in the full release.
Why This Tool Matters
What I found particularly interesting is that the image editing isn’t done by a dedicated image model, but by the multimodal Gemini model. This shows how versatile today’s AI systems have become.
By exploring multiple AI tools, you can access a wider range of features without necessarily paying for premium subscriptions.
Ready to give it a try? Check out Google AI Studio and see what creative edits you can make with just a few text prompts!