Today, AI has reached a somewhat terrifying level, it has been trained so much that it can now create television scripts and can even imitate the art of drawing by anyone who collects its strokes, and this leads us to the fact that they can also, for the benefit of programmers. Theft of information. This brings us to a recent study that warns against uploading photos of minors to the Internet, whether they are acquaintances or one’s own children.
A recent study revealed a disturbing reality: images of children published on the Internet are being used to train AI models without the consent of families. According to Human Rights Watch (HRW)Images of minors, uploaded to the network by their parents or guardians, are used by these models to obtain detailed information about children, violating their privacy and jeopardizing their safety.
It should be noted that AI models can avoid privacy settings and through URL Access sensitive information such as minors’ names and locations, available in data sets. As an example, from a single photo, researchers were able to track two children’s names, ages, and daycares This not only violates privacy, but also exposes children to the possibility of deepfaking or exposing their personal information online
use of The AI training models with unauthorized photos of children raises serious concerns about the privacy and safety of minors. It is imperative that parents and guardians be aware of these risks and exercise caution when sharing their children’s images online. Protecting children’s privacy must be a priority in the digital age, and it is critical that both platforms and laws evolve to address these emerging threats.
Through: Ars Technica
Author’s Note: We certainly aren’t safe uploading content to computers either, so using personal information on networks should be kept as private as possible, only for necessary procedures and nothing else.
