Upload a video to Gemini and quickly discover if it was generated by AI with Google technology.
Key points to note:
– Google is integrating a SynthID watermark checker into the Gemini app specifically for videos.
– This feature can indicate whether a video was created or altered using Google AI.
– The update comes after the introduction of the SynthID watermark validation feature for images in the Gemini app last month.
Over time, distinguishing AI-generated content from genuine photos and videos will become increasingly challenging, highlighting the significance of watermarking and digital metadata. Determining if a video is authentic or AI-produced can be difficult. While there are various online tools that attempt to identify AI-generated content, Google is developing an “imperceptible” watermark technology known as SynthID to aid users in discerning authenticity.
Starting now, Gemini is capable of verifying video authenticity using SynthID within the Gemini app. This video verification tool can identify if submitted content was either created or altered with Google AI in any capacity. Google claims that more than 20 billion pieces of AI-generated content have been marked with SynthID, which digitally watermarks Google AI creations from Veo generations to AI edits in Google Photos.
To utilize the feature, upload a video to the Gemini app and pose a question such as “was this created using Google AI?” Following this, Gemini will analyze the video and provide feedback, including details on which parts of the content (if any) were generated or modified by AI.
Gemini might reply with something similar to “SynthID detected within the audio between 10-20 seconds. No SynthID detected in the visuals,” as noted in Google’s blog post.
The video verification feature does come with certain limitations. For identification via the Gemini app, a video must not exceed 100MB in size and should be 90 seconds or shorter in duration. Moreover, this functionality is limited to spotting videos generated with Google AI tools that utilize the SynthID watermark — it will not disclose information about content created with third-party software.
The new feature in the Gemini app builds on the SynthID photo verification that was launched in November.
“SynthID integrates an invisible digital watermark into an AI-generated image (or video segment),” Google states on the DeepMind website. “The watermark does not alter the quality of the image or video. It is incorporated at the moment of content creation and is designed to endure modifications such as cropping, adding filters, altering frame rates, or lossy compression.”
Now, users of Gemini can effortlessly and quickly verify the authenticity of photos and videos within the Gemini app wherever it is accessible.
