The week has been full of news in Geminithe AI assistant of Google. In addition to the launch of Gemini 3 Pro and Nano Banana Pro in the chatbot, now you can detect if an image has been generated or edited with AI. But there is a nuance, for the moment, and that is that it only works with images created or modified using Google AI.
Gemini uses SynthIDa digital watermarking technology that Google introduced in 2023. SynthID inserts signals, invisible to the human eye, in AI-generated content that remain even after modification.
Google has been testing this technology through its verification portal SynthID Detectoravailable since the beginning of the year for interested journalists, researchers and media professionals. It can detect AI-generated content in different formats and highlight the parts most likely to carry the SynthID watermark.
The company has also integrated SynthID into the generative tool Reimagine of the Magic Editor in the Pixel from Google. So you can quickly confirm whether an image you upload to the Gemini app was created with Google AI by directly asking ‘was this created with Google AI?’ or ‘is this generated by AI?’.
The chatbot will check the SynthID watermark on the uploaded image and use its reasoning capabilities to provide an answer with more context. Later, Google will use its SynthID verification tool to support more content typesincluding audio and video, and will integrate this function into the Seeker and other services.
It’s not just Google that is looking for ways to identify creations made by generative AI. The company is part of the management committee of the C2PA (Coalition for Content Provenance and Authenticity), which also includes Adobe, OpenAI, Goal, Microsoft and other companies and has developed a technology of the same name for this task.
So, C2PA is an open standard that attaches secure metadata to images, audio or video. Adobe has also created its own digital watermarking tool based on C2PA credentials.
Google affirms that its newly released image generation model Nano Banana Pro will insert C2PA metadata into all the content you generate in the Gemini app, Vertex AI and Google Ads. The company has already implemented C2PA credentials in products such as YouTubehe Seeker, Pixel and Photos.
Adding these C2PA credentials to your verification tools will allow users ‘check the original source of content created by models and products that exist outside the Google ecosystem’the company has noted.