Fake news, doctored images, even beauty filters – all of these seem to be more and more common, making it difficult to distinguish truth from lie. But Jigsaw wants to make it easier to debunk all that fakery we see online every day. The Alphabet-owned company has created Assembler, an AI-powered tool that detects image manipulations and combats disinformation.
Deepfake technology, image manipulations and all sorts of filters are becoming more advanced, which is certainly fascinating. But it has a negative side, considering that this advancement is a fertile ground for all kinds of fake news. This is where Assembler comes to the scene. It’s a joint project of Jigsaw and Google Research. Put simply, the platform uses technology trained to identify specific types of manipulation, and it evaluates if and where images may have been altered.
CEO and founder of Jigsaw, Jared Cohen, shared a blog post explaining in more detail how the tool works. Assembler’s detectors can spot specific types of manipulations, from copy-paste to image brightness. The tool scans the image for manipulations, shows where there may have been applied, as well as the probability of these manipulations.
“Additionally, we built two new detectors to test on the platform.The first is the StyleGAN detector to specifically address deepfakes. This detector uses machine learning to differentiate between images of real people from deepfake images produced by the StyleGAN deepfake architecture. Our second model, the ensemble model, is trained using combined signals from each of the individual detectors, allowing it to analyze an image for multiple types of manipulation simultaneously. Because the ensemble model can identify multiple image manipulation types, the results are, on average, more accurate than any individual detector.”
Perhaps you remember a similar tool from Adobe, introduced at the Sneaks event in November 2019. Project About Face also detects image manipulations and how high is their probability. It hasn’t been officially introduced yet (it may not ever be), but it is a good idea, that’s for sure.
As for Assembler, it’s also still in its experimental phase. As Cohen puts it, it’s an “ongoing experiment” considering that “disinformation is a complex problem, and there isn’t any simple technological solution.” Still, in the world of fake news which at the same time spread like wildfire, a solution for combating them is more than necessary. Therefore, I hope that Assembler will see the light of day soon.