Deepfake Forensics Workshop (DFF-2026): Detection, Attribution, Recognition, and Adversarial Challenges in the Era of AI-Generated Media
Abstract
The rapid evolution of generative artificial intelligence is profoundly transforming the multimedia ecosystem, enabling the creation of highly realistic synthetic and manipulated content while simultaneously introducing unprecedented challenges for the authenticity, reliability, and trustworthiness of digital media. As generative technologies continue to evolve, multimedia forensics is no longer limited to deepfake detection, but is progressively expanding toward a broader research landscape encompassing source attribution, content provenance, explainability, adversarial robustness, multimodal forensic analysis, and evaluation under realistic deployment conditions. This evolution requires forensic methodologies capable of operating across heterogeneous generation pipelines while remaining reliable, interpretable, and applicable in real-world scenarios. The 2nd Deepfake Forensics Workshop (DFF-2026) provides an interdisciplinary forum for researchers and practitioners from multimedia, computer vision, machine learning, cybersecurity, and digital forensics to discuss these emerging challenges and foster the development of next-generation forensic solutions. Building upon the success of the inaugural edition, the second workshop broadens its scientific scope to address the entire lifecycle of AI-generated multimedia, from content generation and manipulation to forensic analysis, authentication, provenance verification, and trustworthy deployment. This paper reviews the scientific directions that emerged from the first edition of the workshop, discusses how they have shaped the evolution of the DFF initiative, and presents the objectives and research themes that characterize DFF-2026, highlighting its role as a meeting point for the multimedia forensics community in the era of generative AI. Further information about the workshop is available at https://iplab.dmi.unict.it/mfs/acm-dff-ws-2026/.