gen‑ai.news
← Back
Image

Google will now allow users to remove visible watermark from its AI generations

Google has updated its AI image and video generation tools to let users opt out of the visible watermark that was previously stamped on all outputs. The watermark, which displayed a visible indicator that content was produced by one of Google's generative AI systems, will now be toggleable - meaning creators can turn it off if they prefer a cleaner result for their work.

Crucially, disabling the visible watermark does not affect SynthID, Google's invisible watermarking technology. SynthID embeds an imperceptible signal directly into the pixels of an image or the frames of a video, allowing the content to be identified as AI-generated even after the visible branding has been removed. This distinction matters because SynthID operates at a technical level that is not accessible to end users - it persists through many common edits and cannot be switched off through the same setting.

The move reflects a tension that has been building across the generative AI industry. Visible watermarks are a straightforward way to signal AI origin to viewers, but many creators and professionals have argued they limit the practical usefulness of generated content, particularly in commercial or editorial contexts. By separating the visible branding from the underlying provenance signal, Google is attempting to give users more control without fully abandoning the infrastructure it has built for AI content identification.

SynthID was first introduced by Google DeepMind in 2023 and has since been extended beyond images to cover audio, video, and text. Its persistence even when visible markers are removed positions it as the more durable layer of Google's approach to AI transparency - one that does not depend on user cooperation to function. Whether that is sufficient for regulators and platforms increasingly focused on AI content disclosure remains an open question, but the separation of visible and invisible watermarking gives Google some flexibility as those standards continue to develop.

Enjoy this story? Get the next one in your inbox.

Twice a week: the most important stories in generative image and video AI, distilled into a 2-minute read.

Free. Unsubscribe any time. No spam, ever.

Your next read

No image
Image

With Dazzle, Marissa Mayer bets your camera roll has more info on your life than your inbox

Marissa Mayer, the former Yahoo CEO, is launching Dazzle, an AI personal assistant that draws entirely on a user's camera roll rather than email or calendar data. The premise is that photos contain a richer, more candid record of a person's life than a inbox ever could. The app uses generative AI to interpret that visual history and surface useful, context-aware assistance.

Italian Politician Sparks AI Scandal With Doctored Photo That ‘Undressed’ Female Lawmaker
Image

Italian Politician Sparks AI Scandal With Doctored Photo That ‘Undressed’ Female Lawmaker

An Italian politician has drawn widespread condemnation after sharing an AI-manipulated image that appeared to undress a female colleague. The incident arrives months after Prime Minister Giorgia Meloni herself shared a deepfake nude of her own likeness to publicly highlight the dangers of such technology. The case is reigniting debate across Europe about legislative responses to non-consensual AI-generated imagery.

Humans Are Reading Copilot Prompts — And They're Horrified
Image

Humans Are Reading Copilot Prompts — And They're Horrified

Internal documents obtained by 404 Media reveal that human contractors are reviewing prompts and uploaded images submitted by Microsoft Copilot users. The reviewers are reportedly exposed to large volumes of explicit image generation requests, raising questions about privacy, consent, and working conditions in AI data pipelines.