Google subtly revealed last week that it would be incorporating a visible watermark into videos produced by AI with its new Veo 3 model.
Additionally, you may be able to spot it if you pay great attention while browsing through your social media feeds.
Videos that Google made to advertise the Veo 3 launch in the UK and other countries show the watermark.
Josh Woodward, a vice president with Google Labs and Google Gemini, disclosed the shift in an X thread.
The business reportedly placed a watermark to all Veo videos, with the exception of those created by users with a Google AI Ultra plan using Google’s Flow tool. The new watermark is in addition to the SynthID detector, which was just made available to early testers but is not yet widely accessible, and the invisible SynthID watermark that is already included in all of Google’s AI-generated content.
According to Josh Woodward, VP of Google Labs and Gemini, the visible watermark “is a first step as we work to make our SynthID Detector available to more people in parallel,” as said in his X post.
Given that it can also produce realistic audio and conversation, Veo 3, Google’s new AI video model, has attracted a lot of attention in the weeks since it was shown at Google I/O 2025. While there are many videos of animals acting like people on the internet, they are not the only examples. More commonplace clips, such as influencer advertisements, fake news pieces, unboxing videos, and man-on-the-street interviews, have also been produced using Veo 3.
If you look closely, you can see obvious indications of artificial intelligence, such as excessively smooth skin and inaccurate backdrop artifacts. However, if you’re passively doomscrolling, you might not consider verifying whether the kangaroo carrying a travel ticket in a casual manner is genuine or not. An example that is comparatively innocuous is when an AI-generated kangaroo tricks people. But according to AI specialists Mashable spoke with for this article, Veo 3’s accessibility and realism provide a new degree of risk for the propagation of false information.
Theoretically, the new watermark should lessen those dangers. The visible watermark isn’t really visible, which is the main issue. A “Veo” watermark in a light white hue can be seen in the lower right-hand corner of a video that was created with Veo 3. Can you see it?
Hany Farid, a specialist in digital forensics, stated that most users who are scrolling through their social media feed at a rapid pace are unlikely to see this little watermark. Yes, we were searching for it, and it took us a few seconds to locate it. Users might not notice the watermark unless they know to search for it, particularly if they are viewing content on a mobile device.
“We’re committed to developing AI responsibly and we have clear policies to protect users from harm and govern the use of our AI tools,” a Google representative emailed Mashable. SynthID watermarks are included in every Google AI-generated content, and we also add a visible watermark to Veo videos.
Negar Kamali, a researcher at Kellogg School of Management who studies people’s capacity to recognize AI-generated content, stated, “People are familiar with prominent watermarks like Getty Images, but this one is very small.” “So either the watermark needs to be more noticeable, or platforms that host images could include a note beside the image — something like ‘Check for a watermark to verify whether the image is AI-generated,'” Kamali said. “Over time, people could learn to look for it.”
Visible watermarks aren’t a foolproof solution, either. Watermarked videos are easily cropped or manipulated, according to both Farid and Kamali. Farid, a professor at the UC Berkeley School of Information, stated, “none of these small — visible — watermarks in images or video are sufficient because they are easy to remove.”
But he pointed out that the invisible stamp created by Google, SynthID, “is quite resilient and difficult to remove.” According to Farid, “The downside is that the average user can’t see this [SynthID watermark] without a watermark reader so the goal now is to make it easier for the consumer to know if a piece of content contains this type of watermark.”
