Google Gemini AI Faces Criticism Over Watermark Removal Capabilities

The experimental Gemini 2.0 Flash model can remove watermarks from images, raising ethical and legal concerns about copyright violations.

Google's Gemini 2.0 Flash AI model, accessible via AI Studio, can effectively remove watermarks from images, including complex and stylized ones.
Users have demonstrated the tool's ability to remove watermarks on social media, sparking concerns over potential misuse and copyright infringement.
Unlike other AI models like OpenAI's GPT-4o and Anthropic's Claude, Gemini does not warn users about the ethical or legal implications of watermark removal.
The tool also reconstructs and upscales images during the watermark removal process, achieving high-quality results that could facilitate unauthorized use.
Google has not yet responded to inquiries about the lack of guardrails in Gemini 2.0 Flash, which is currently labeled as experimental and not for production use.