Google has updated the Gemini app to detect whether an image was generated or edited using AI. However, the catch here is that the new feature currently works for images generated or edited with Google AI.
The force behind Gemini"s AI image detector is SynthID, a digital watermarking technology Google introduced in 2023. SynthID embeds imperceptible signals into AI-generated content that remain intact after it"s modified.
Google has been testing the technology through its SynthID Detector verification portal, which was made available to interested journalists, researchers, and media professionals earlier this year. It can spot AI-generated content across different modalities and highlight the parts most likely to be watermarked with SynthID.
The search giant also added SynthID support to the Reimagine generative AI tool in Magic Editor. That said, you can quickly confirm whether an image you uploaded in the Gemini app was generated using Google AI by asking questions: "Was this created with Google AI?" or "Is this AI-generated?"
The chatbot will verify the SynthID watermark in the image you uploaded and use its reasoning capabilities to provide a response that adds more context to that image. Moving forward, Google will use its SynthID verification tool to support more content types, including audio and video, and bake the feature into Google Search and other places.
It"s not just Google; the rising threat of AI-generated content that misleads and deceives users has woken up other tech companies. Google is a steering committee member of the Coalition for Content Provenance and Authenticity (C2PA), which also includes Adobe, OpenAI, Meta, Microsoft, and others.
While SynthID is Google"s proprietary effort, C2PA is an open standard that attaches secure metadata to images, audio, or video. Adobe also developed its own digital watermarking tool based on C2PA credentials.
Google said its freshly baked Nano Banana Pro image-generation model will embed C2PA metadata in its generated content across the Gemini app, Vertex AI, and Google Ads. Google has implemented C2PA credentials across products like YouTube, Search, Pixel, and Photos.
Adding C2PA content credentials to its verification tools in the future will allow users to "check the original source of content created by models and products that exist outside of Google’s ecosystem," the company said.