US Patent 12062105 Utilizing multiple stacked machine learning models to detect deepfake content

A device may receive training data for training a first machine learning model, a second machine learning model, and a third machine learning model and may train the first machine learning model, the second machine learning model, and the third machine learning model with the training data. The device may receive input content and may process the input content, with the first machine learning model, the second machine learning model, and the third machine learning model, to generate a first model result, a second model result, and a third model result, respectively. The device may process the first model result, the second model result, and the third model result, with an aggregation model, to generate a detection result indicating whether the particular input content is a deepfake or is real and may perform one or more actions based on the detection result.