November 21st.GoogleRight now Gemini USERS CAN EASILY DETECT PARTIAL AI GENERATION. THIS IS THE DAYUsers can use Gemini applications to judge whether a picture is created or edited by the Google AI tool, just ask Gemini, "Is this AI generated?"。
According to 1AI, while initially limited to images, Google has indicated that the video and audio content validation function will be launched “promptly” and will be extended to Gemini applications in the future, including integration into Google Search。
More important expansion will take place in the future:GOOGLE PLANS TO EXTEND THE CERTIFICATION CAPACITY TO SUPPORT INDUSTRY-WIDE C2PA CONTENT CERTIFICATE STANDARDSI don't know. The current image validation is based only on the invisible AI watermark technology developed by Google; once extended to the C2PA standard, more types of AI tools and creative software (including Sora of OpenAI) will be identified。
In addition, Google announced today that the images generated by its newly released Nano Banana Pro model would be embedded in C2PA metadata. Previously, TikTok had confirmed that C2PA metadata would be used in its invisible watermarking programme for the content generated by its AI。
While manual content validation in Gemini is a practical development, C2PA vouchers and other watermarks, such as SynthID, will only work when social media platforms are better able to automatically mark artificially generated content, rather than putting the onus on users。