Whereas Google’s AI-powered Magic Editor characteristic is little doubt spectacular, there’s little doubt that any AI-based enhancing software program nonetheless runs the chance of getting used for extra less-than-noble causes, particularly given how AI-related scams and misinformation are prevalent these days. With that in thoughts, Google not too long ago introduced that it’s implementing a digital watermark system utilizing SynthID’s know-how.
For these unfamiliar, SynthID is a method of embedding digital watermarks inside AI-generated audio, textual content and even video, which will probably be used to mark photographs edited utilizing Magic Editor’s “Reimagine” perform. It’s truly already utilized in AI-generated photos created by way of Google’s Imagen AI mannequin, permitting customers to tell apart AI-made content material.
READ: Google Lastly Provides a Fundamental Function to its Pictures App
Google does make clear that there could be some instances the place SynthID might need hassle detecting AI edits, particularly in the event that they’re small – for instance, for those who change the color of a rose petal within the background of a photograph, the system won’t be capable to determine and label it as an AI-edited picture. Google provides that it’s going to proceed working with person suggestions and transparency concerning AI edits made in Google Pictures.
