Experts at Michigan State University and Facebook have developed a reverse-engineering research method to detect and attribute deepfakes. The method relies on uncovering the unique patterns behind the artificial intelligence (AI) model used to generate a single deepfake image. Initially, a Fingerprint Estimation Network is used to process an image and look for hidden patterns that indicate the photo is a computer-generated fake. Then a Parsing Network analyses these hidden patterns to predict the number of layers in the deepfake neural network and how they might be connected. Ultimately, the output of the Fingerprint Estimation Network is fed into a binary clarifier to determine if a picture was computer-generated. The model is intended to assist researchers and practitioners in better investigating cases of coordinated disinformation using deepfakes.