Google AI watermarking system allegedly reverse-engineered

A software developer claims to have successfully reverse-engineered Google DeepMind's SynthID system, which is designed to watermark AI-generated images. The developer, known as Aloshdenny, has published their findings on GitHub, detailing a process that allegedly requires only a set of AI-generated images and signal processing techniques, without needing neural networks or proprietary access. This development raises questions about the robustness of current AI watermarking technologies and their ability to prevent the misuse or misattribution of synthetic media. Google has reportedly disputed the claim, stating that the system has not been compromised.