The new tool relies on “SynthID” technology, an invisible digital watermarking technology developed by Google in 2023, which works by embedding a digital watermark in content created by the company’s AI models. This watermark is detectable even if the content is cut, copied, or modified.
Google explained in an official statement that the launch of the video verification tool is part of a larger plan to expand content transparency tools, especially with the rapid development of AI models capable of producing highly realistic images and videos, which increases the risk of misinformation and the spread of misleading information. Last November, the company added a similar tool to detect images created by AI within Gemini, before expanding “SynthID” capabilities to include video, audio, and text.
Users can upload a video to the Gemini application or website and then ask the chatbot directly: “Was this video created or modified using artificial intelligence?”
Gemini will then analyze the audio and visual tracks separately to search for the “SynthID” watermark, and if found, will provide a detailed response indicating where it is located, such as pointing to a specific time in the audio or video, while specifying whether the modification included one or both elements.
This update comes at a time of growing global concerns about the use of artificial intelligence in manipulating visual and audio content, whether for purposes of misinformation or fraud. Through “SynthID,” Google aims to provide a reliable technical means to help users, journalists, and content creators verify the source of digital materials, without affecting content quality or viewing experience. (اليوم السابع)
