The creation and manipulation of multimedia content has always been a central topic in the multimedia community. For decades, images, videos, and audio recordings have been edited, enhanced, or tampered with using both professional tools and consumer software. However, recent advances in artificial intelligence and generative modeling have radically changed the scale, realism, and accessibility of media manipulation. Multimedia content can be altered through a wide spectrum of techniques: from classical post-processing and compositing, to fully synthetic generation, to hybrid editing pipelines combining traditional tools and AI-based methods. In many cases, highly realistic manipulations can be produced with minimal effort, sometimes even through high-level semantic instructions such as text prompts. As a result, the barrier between authentic and manipulated content is becoming increasingly blurred. This technological change is already having significant social, legal, and security implications. Manipulated and synthetic media are often used for malicious purposes: disinformation and propaganda; non-consensual and abusive content (e.g., deepfake pornography); fraud and identity theft; reputational attacks; and potentially the fabrication or alteration of digital evidence. In this context, the role of multimedia forensics is becoming both more critical and more challenging. The problem is no longer limited to detecting low-quality or visually inconsistent forgeries, but instead concerns the reliable analysis of highly realistic and semantically coherent manipulations. Modern forensic methods are required to detect and localize manipulations across heterogeneous editing and generation pipelines, and to attribute synthetic or manipulated content to specific models, tools, or generation processes. At the same time, forensic techniques must remain effective in a technological landscape where generation and editing methods are rapidly evolving and continuously optimized to remove detectable traces. Therefore, there is a growing need for new tools that are not only accurate, but also robust, interpretable, and suitable for use in real investigative and judicial contexts. At the same time, the field is increasingly challenged by adversarial machine learning and counter-forensic strategies. Existing forensic systems often suffer from limited generalization and are vulnerable to targeted attacks designed to evade detection or mislead attribution. The rapid and continuous evolution of generation and manipulation technologies demonstrates that ad-hoc or artifact-specific detectors and related solutions are no longer sufficient, calling for more general and robust approaches.% This workshop aims to consolidate and advance research on the forensic analysis of manipulated and AI-generated multimedia content, including images, videos, and audio, and addressing the full spectrum of challenges ranging from detection and localization to attribution, robustness, explainability, and legal implications. The workshop is designed to cover both the analysis of existing manipulation and generation techniques and the study of new synthesis and editing methods, with the goal of understanding how different manipulation pipelines affect forensic traces and the reliability of detection systems.