Google is enhancing its Gemini app by integrating AI content detection through its invisible “SynthID” watermarks, enabling users to ascertain if an image was generated by AI. This feature allows Gemini to identify these watermarks automatically embedded by Google’s AI models, providing a reliable method of detection. Users can upload images and inquire, “Is this image generated with AI?” Gemini will analyze and report findings related to SynthID watermarks or other AI indicators. Although effective, the limitation is that SynthID only applies to Google AI models, meaning content from other platforms, like Midjourney, remains undetectable. Gemini’s detection capabilities also extend to spotting visual artifacts and clues indicative of AI generation. While this feature marks progress in verifying content validity, it doesn’t completely guard against deceptive usage. Google aims to extend SynthID’s application beyond images to video, audio, and text, fostering a culture of content verification among users.
Source link
