Google announced on Friday that it’s going to now permit customers to take away a visual watermark from its AI generations, together with photos, movies, and songs. The corporate specified that this gained’t have an effect on the invisible SynthID watermark and C2PA standard-related metadata.
Josh Woodward, the corporate’s VP for Gemini, said in a post on X that this toggle can be obtainable for Nano Banana, Omni, and Lyria fashions. He specified that the setting to show the seen watermark off can be obtainable in Gemini and Google’s video editor, Flow, with help for Search coming quickly.
The change displays an evolving strategy to labeling AI-generated media, as seen watermarks can typically make content material much less helpful for skilled and artistic work, however the necessity to establish AI-generated content material stays.
“We’re putting a stability right here between inventive management and security: whereas the seen watermarks are actually elective, invisible SynthID watermarks and C2PA metadata are nonetheless getting used for transparency. So you’ll be able to nonetheless use Gemini or Search to see if a picture was AI-generated,” he wrote.
The function is rolling out within the coming days, and as soon as obtainable, customers will be capable to go to Settings > Media Watermark to show the seen mark on or off.
Google can be open sourcing a brand new library known as Credentio to let builders embed a neighborhood validation mechanism of their apps.
This comes after Anthropic’s hotly debated move to incorporate a watermark in text and files generated by Claude to adjust to EU rules.
