The advancements in AI are endowing works generated with this technology with an ever-greater degree of realism, which can present a significant challenge when it comes to distinguishing between reality and artificial content. To assist users in this regard, Google has integrated within Gemini the capability to detect whether a video has been edited or created using any of Google’s AIs.
This feature has been deployed in all languages and countries supported by the Gemini application.
Whenever we have introduced any of Google’s generative AI tools, such as, for example, VEO2 or Nano Banana Pro, we have also taken the opportunity to inform you about SynthID, the invisible digital watermark embedded in all content generated by Google’s AI, which serves to identify it as such.
Now, Gemini possesses the capability to identify the invisible SynthID watermark in images as well as in audio and video tracks, allowing any user to determine if an image or video was created, either wholly or partially, with Google’s AI. For example, Gemini could inform you: “SynthID detected in audio between 10 and 20 seconds. No SynthID detected in images.”
It is important to emphasize that files you upload to Gemini may have a maximum size of 100MB and a duration of 90 seconds. If the files exceed these limits in size or duration, the AI will not be able to process them.
To verify content, you only need to follow these simple steps:
In our editorial office, we have tested this feature, and Gemini was not only able to detect the SynthID watermark in the image but was also able to apply its reasoning to identify additional clues indicating that the content was AI-generated.
Photo: generated with Nano Banana Pro (Gemini)
Your email address will not be published. Required fields are marked *
Δ