"Developed in 2023, SynthID has been part of Google's various image-generating AI models for a couple of years now. The watermarking system is designed to be imperceptible to humans and still detectable when images are cropped or modified, according to Google. SynthID is open source, and Google has scored a few high-profile partners in Hugging Face and Nvidia, but aside from those examples, Gemini's SynthID Detector won't be able to actually determine whether an image is AI-created with any sort of certainty."
"In our testing of AI images generated via ChatGPT, Gemini wasn't reliably able to tell if the picture was AI-generated or not, but it did reason correctly a few times based on small details that give AI-generated content away. Images created using Gemini, on the other hand were all flagged as containing SynthID watermarks. Instead of relying on watermarking like Google, ChatGPT relies on a metadata system developed by the Coalition for Content Provenance and Authenticity (C2PA) to mark its content as AI-generated."
Google released SynthID Detector in the Gemini app and on the web, enabling users to upload images and ask whether they were created or modified using AI. The detector can only recognize images tagged with Google-made SynthID watermarks embedded by Gemini. SynthID is imperceptible to humans yet detectable after cropping or modification. SynthID is open source and has partners such as Hugging Face and Nvidia. The detector cannot determine AI origin with certainty for images lacking SynthID tags. Gemini struggled to reliably identify images generated by ChatGPT, while Gemini-created images were consistently flagged. Google plans to add C2PA support to Gemini.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]