St Hilda’s Anglican School for Girls removes students’ photos from social media to prevent deepfake abuse

An elite Perth girls’ school is erasing identifiable images of students from its public social media sites to prevent them from being downloaded and altered with artificial intelligence tools to create harmful deepfakes. St Hilda’s Anglican School for Girls has revealed that from this year, it will post only non-identifiable photos such as silhouettes, backs of heads or distant group shots to guard against the growing threat. “Overseas, there have been serious cases where images and names taken from a school’s social media have been altered into harmful deepfakes and used for ransom threats,” the Mosman Park school told parents. “Once an image is publicly posted, it can be downloaded, manipulated and shared without our knowledge or consent.” St Hilda’s principal Fiona Johnston said she was not aware of any incidents involving deepfakes or misuse of its students’ images. “We felt it was important to act before harm occurs, rather than in response to an incident. “Once images are shared publicly, they can be downloaded, altered or redistributed without a school or family ever knowing.” Ms Johnston said the digital environment had “changed significantly” in recent years, with the rapid advancement of AI, image manipulation and deepfake technology. “These developments present new risks for children that schools did not previously have to consider,” she said. “To our knowledge, St Hilda’s is the first school in Australia to publicly commit to removing clearly identifiable student images from public social media channels as part of a formal strategy.” Students’ images would still be available to parents on a password-protected site run by media management platform Pixevety. The move comes after Elon Musk’s X and AI chatbot Grok sparked international outrage by allowing users to digitally “strip” photos of women and children to create sexualised images and the rise of other “nudifying” apps. Perth cybersafety expert Paul Litherland said a growing number of schools had contacted him for advice on the best approach to using student photos on social media sites. “I think it’s a good start to at least try to minimise the availability of the images of our students on public systems,” he said. Mr Litherland said he believed the trend to use AI to create harmful deepfakes would continue to evolve. “This is going to be our next big wave in regards to education,” he said. “And, I’m hoping, a legislative response as well.” While Australia had clamped down on children’s social media use, he believed there was not enough moderation for AI tools. He urged parents, educators and organisations to remember that anything could be pulled off line and manipulated. The eSafety Commission last year issued a warning to schools about the rapid rise of explicit deepfake images being used to humiliate, bully and sexually extort children in the schoolyard and beyond.