This New AI Model Let’s You Remove Watermarks From Image

Things are about to change with this new AI model that removes watermarks from images. We all have struggled to get those beautiful copyrighted free images from the internet and most probably settled for something less as not all of those came free. The popular websites like pixabay, freepik etc comes with a price if you really want that good quality copyright and watermark free image.

Well, as we are waking up every single day to new AI models from tech giants, this time it’s the remove watermarks from images capability that had us impressed – with Google’s Gemini 2.0 Flash.

What is Google Gemini 2.0 Flash?

Google has been rolling out Gemini and upgrading it’s capabilities competing against Microsoft’s Co-pilot and OpenAI’s GPT 4. With the new update, now users can remove watermarks from images.

Initially designed for image generation and editing, Gemini 2.0 Flash was introduced as an “experimental” feature in Google AI Studio, a platform which is mostly.

This AI model can generate visiuals based on user’s prompts and also remove watermarks from images that we find on platforms like Getty Images and Shutterstock.

How Good Is It at Removing Watermarks?

Well, it’s actually good. While the other AI models struggles leaving behind awkward smudges, this ai model that removes watermarks from images – Gemini 2.0 Flash fill those gaps quite seamlessly and polished.

Although it replaces with it’s own tiny mark at bottom left corner with a subtle ‘edited with AI’ which isn’t really hard to remove that too with other techniques.

Gone Too Far?

While it may look like a cool tool for your personal projects, but deep down there are genuine ethical and legal concerns about all this.

• Copyright Violations – Watermarks aren’t just a stamp. They exist to protect the intellectual property of the creator and removing them without their consent is legally inappropriate. 
• Transparency issues – When watermark is removed, it’s even hard to tell if the image is coming from a genuine creator or an altered content.
• Potential Misuse – Deepfakes and misinformation is rising more than ever with new AI models, now with something that can turn copyright image into a “free” material could add to that chaos.

Can AI outsmart AI?

Previously Google has tried to address the issue of deepfakes and copyright with SynthID which is a technology from Google DeepMind that watermarks and identifies AI-generated content by embedding digital watermarks directly into AI-generated images, audio, text or video. But as capable as Gemini is at removing watermarks, there’s a concern it could squash SynthID too. This looks like a battle of AI against AI which we have looked forward to as the way the world is heading.

What happens next?

For now, it looks like some cool tech to play with around with, although when it becomes mainstream – it could complicate how we handle copyright content online. 

Could this be a shift on how things have worked all this while or time to a accept the new realities of the rapidly advancing AI world?

Would love to hear your thoughts on this, drop a comment and let us know!