The field of synthetic media detection and forensics is rapidly evolving, driven by the increasing sophistication of generative AI technologies. Recent developments have focused on improving the detection of AI-generated images and videos, as well as enhancing the robustness of forensic tools against various types of image and audio manipulations. Notable advancements include the development of autonomous and self-adaptive systems for synthetic media detection and attribution, which can identify and incorporate novel generators without human intervention. Additionally, innovative approaches such as frequency-aware learning and wavelet prompt tuning have shown significant improvements in detecting deepfakes and other types of synthetic media. These advancements have the potential to mitigate the risks associated with synthetic media, including disinformation, fraud, and other malicious applications. Noteworthy papers include:

• The paper introducing a multi-feature fusion framework for robust AI-synthesized image detection, which achieved exceptional generalization performance across diverse GenAI systems.
• The work presenting AnomalyHybrid, a domain-agnostic framework for general anomaly detection, which demonstrated superior performance in generating authentic and diverse anomalies.
• The study proposing a wavelet prompt tuning method for enhanced auditory perception in all-type deepfake audio detection, which achieved the best performance with an average EER of 3.58% across all evaluation sets.