AI Training Dataset for Detecting Nudity Allegedly Found to Contain CSAM Images of Identified Victims

The NudeNet dataset, containing more than 700,000 images scraped from the internet, was used to train an AI image classifier designed to automatically detect nudity in images. The dataset was made available for download from Academic Torrents in June 2019. The Canadian Centre for Child Protection (C3P) discovered that the dataset contained more than 120 images of identified or known victims of child sexual abuse material (CSAM), including nearly 70 images focused on the genital or anal area of children who are confirmed or appear to be pre-pubescent. Some images depicted sexual or abusive acts involving children and teenagers such as fellatio or penile-vaginal penetration. More than 250 academic works either cited or used the NudeNet dataset since its availability. A non-exhaustive review of 50 academic projects found 13 made use of the NudeNet dataset and 29 relied on the NudeNet classifier or model. People and organizations that downloaded the dataset would have no way of knowing it contained CSAM unless they went looking for it, but having those images on their machines would be technically criminal. Academic Torrents removed the dataset after C3P issued a removal notice to its administrators.