Generative AI image outputs are becoming more convincing, leading to concerns about distinguishing between real and fabricated content. One way to identify AI-generated images is through the embedded watermarks, although looking for the visible Gemini logo is not always a reliable method. This is especially true now that Google is giving users the option to remove said watermark.
In an X post, Google Labs Vice President Josh Woodward revealed that the company is making the visible watermark optional via a new setting. This setting, which lets users toggle visible watermarks on or off, is available for Gemini and Flow and will expand to AI Mode in Search.
As mentioned in Woodward’s post, the setting applies to images created with Nano Banana, as well as videos generated via Omni, and songs made with Lyria. He also noted that there is an exception: the watermarks will remain in countries where it’s “required by law” to keep them.
It is worth noting that the setting only applies to visible watermarks, which are arguably already easy to remove to begin with. On the other hand, the invisible SynthID and C2PA metadata remain embedded in the content. These labels, while unseen to the naked eye, can still be identified with Gemini or Search.
According to Woodward, the change is meant to strike a balance between creative control and safety. With this setting, those who rely on the Gemini to generate images won’t have to go through all the trouble to remove that pesky logo before sharing them. On the flip side, untrained eyes will have to work harder to figure out if what they’re looking at is a real photograph.
That said, Google is not the first to make the obvious branding optional. OpenAI and Meta already allow users to generate images without a visible watermark. Of course, the unseen labels still remain attached to the content.
(Source: Josh Woodward via X)
