Google's Gemini 2.0 AI Sparks Controversy with Watermark Removal Feature
March 17, 2025
Google's latest AI model, Gemini 2.0 Flash, has sparked controversy as it can remove watermarks from images, raising significant concerns about copyright infringement.
Users on platforms such as Reddit and X have showcased this capability by posting examples of stock images with removed watermarks, highlighting the potential for misuse.
While the watermark removal feature is labeled as 'experimental' and intended for developer use only, its effectiveness has led to worries about its implications for content creators.
The core issue lies not in the technology itself but in its potential misuse; thus, Google bears the responsibility to implement safeguards against such practices.
In response to the backlash, Google has reiterated that using their AI tools for copyright infringement violates their terms of service and they are closely monitoring user feedback.
This situation reflects a broader trend, as seen in China's recent regulations requiring AI-generated content to be labeled, indicating a global push for stricter oversight of AI technologies.
The ongoing debate underscores the challenges of balancing AI advancements with ethical and legal standards in content creation, particularly as AI technology evolves.
Google's SynthID, developed by DeepMind, aims to address these concerns by embedding an imperceptible digital watermark into AI-generated images, enhancing transparency.
As copyright holders express their concerns, it is anticipated that Google may implement updates to restrict the watermark removal functionality in future versions of its AI tools.
The potential for misuse is significant, as the model allows easy access to stock images, raising alarms about the impact on creators' livelihoods and the prevalence of deepfakes.
The legality of generating AI images of celebrities adds another layer of complexity, involving rights of publicity and concerns over deepfakes and defamation.
As AI technology continues to advance, there is an urgent need for countries to update existing laws or create new legislation to protect copyright holders and ensure responsible AI use.
Summary based on 36 sources
Get a daily email with more Tech stories
Sources

TechCrunch • Mar 16, 2025
People are using Google's new AI model to remove watermarks from images | TechCrunch
TechCrunch • Mar 17, 2025
People are using Google's new AI model to remove watermarks from images | TechCrunch
The Verge • Mar 17, 2025
Google’s Gemini AI is really good at watermark removal
Mashable • Mar 17, 2025
Google's new AI model is being used to remove image watermarks