AI image generators can be enjoyable to interact with, but they have a drawback - their generative nature. These algorithms gather information from the internet and produce output influenced by those sources. This can be problematic as it may result in the AI learning from and replicating artwork or photography without proper compensation to the artist or photographer. Furthermore, people who appear in photographs might not appreciate AI altering their likeness in strange or unauthorized ways. To address this issue, MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) created a tool called "PhotoGuard" to safeguard images against malicious editing.

The key to PhotoGuard's functionality lies in the manipulation of pixels, making subtle changes that are imperceptible to humans but significantly disrupt the algorithms. Two methods have been developed by MIT, as described by Hadi Salman, the lead author and a doctoral student at the institute. Salman also proposed that AI model manufacturers could provide application programming interfaces (APIs) to protect and "immunize" the photos of other individuals. This suggestion holds merit, particularly in light of numerous lawsuits in which AI models have been trained on the works of creators without their consent, leading to legal disputes like the case involving Getty Images and Stability AI, where over 12 million images were allegedly copied without permission or compensation. Therefore, exploring avenues to safeguard creative content and respect the rights of authors, musicians, and other creators is crucial.