eWEEK content material and product suggestions are editorially unbiased. We could generate income if you click on on hyperlinks to our companions. Learn More.
Google’s latest AI mannequin, Gemini 2.0 Flash, is exceptionally good at eradicating watermarks from copyrighted photographs together with Getty Images, in accordance with a number of customers. However, AI watermark removing does increase copyright considerations.
Gemini 2.0 Flash is a developer-facing AI instrument
Gemini 2.0 Flash is free to make use of, which additional will increase the AI mannequin’s attraction; it’s only accessible by way of Google’s developer-facing AI instruments, reminiscent of Google AI Studio.
To alter a preexisting picture with the Gemini mannequin, or to create a brand new picture, customers should choose “Gemini 2.0 Flash (Image Generation) Experimental.” Users should additionally make sure the “Images and text” output kind is chosen, relatively than simply textual content.
Gemini 2.0 Flash fills in eliminated watermarks with good AI
The Gemini 2.0 Flash mannequin doesn’t simply remove the watermark — it additionally fills in that area to mix the watermark’s removing. All generative AI instruments try this course of to some extent, however Gemini seems to be extra superior in filling in gaps seamlessly. Gemini does add its personal delicate watermark to the ultimate picture: a small blue diamond within the backside left hand nook.
In addition, the AI mannequin generally struggles to utterly take away clear and/or very giant watermarks.
AI watermark removing raises copyright considerations
The Gemini 2.0 Flash mannequin is labeled “experimental” and “not for production use.” The AI mannequin’s talents recommend it at the moment has few guardrails to restrict its photographs creation and modification talents; that is particularly notable since a number of different generative AI fashions, reminiscent of Anthropic’s Claude 3.7 Sonnet and OpenAI’s GPT-4o, don’t permit watermark removing.
Removing a watermark with out permission from the unique picture proprietor is taken into account unlawful underneath U.S. copyright legal guidelines, besides in uncommon instances. When requested to touch upon the scenario by TechCrunch, a Google spokesperson offered the next assertion: “Using Google’s generative AI tools to engage in copyright infringement is a violation of our terms of service. As with all experimental releases, we’re monitoring closely and listening for developer feedback.”
