AI Training on Children's Photos Raises Privacy Concerns

Human Rights Watch reports widespread misuse of Australian children's images in AI datasets without consent.

Human Rights Watch found 190 photos of Australian children in the LAION-5B dataset, used by AI tools like Stable Diffusion and Midjourney.
Photos included identifiable information such as names, ages, and locations, posing significant privacy risks.
Indigenous children are particularly vulnerable due to cultural restrictions on image reproduction.
Current AI models cannot forget data they've been trained on, even if removed from datasets later.
Australia's upcoming Privacy Act reforms may introduce new protections for children's online privacy.