AI-generated media encompasses a broad spectrum of content, including images, videos, audio, and text, produced with AI technologies. Recent advances in AI models offer numerous benefits for humans in various aspects of life and work. Specifically, the benefits of AI-generated media span a wide range of fields, from enhancing education and healthcare to boosting creativity, optimizing business processes, and improving accessibility. However, multimedia generated by AI models also carries significant risks of misuse and societal upheaval. Specifically, these models can be employed to craft convincing fake news and deepfakes, thereby challenging the information accuracy and public trust.

Despite substantial progress in generative modeling (e.g., GANs, diffusion models, and large language models), current research often lacks a coordinated focus on responsible design, fair deployment, and societal alignment of such technologies. Moreover, the tools for detection, provenance tracing, watermarking, and forensic analysis often lag behind the pace of generation techniques, creating asymmetries that adversaries can exploit. This Research Topic aims to address the following core problem: How can we ensure the responsible and secure development, detection, and deployment of AI-generated media in alignment with societal values and public trust? To address this, we seek contributions that:

• Analyze emerging trends and societal impacts of generative AI across domains such as elections, education, journalism and healthcare.
• Propose novel detection and authentication techniques for AI-generated content across modalities.
• Investigate robustness, fairness, and generalization of detection models under distribution shifts.
• Develop frameworks for watermarking, provenance tracing, and secure media attribution.
• Explore regulatory, ethical, and human-centered considerations in deploying generative technologies.

We welcome original research articles, reviews, and case studies that advance this field, including but not limited to:

AI-Generated Media:
• Foundation Models, Diffusion Models, Large Multimodal Models (Agents) in Multimedia
• Decentralized Foundation Models from Small Data
• Media Generation with multimodal Large Language Models
• Causal and Mechanistic Explanations of Large Language Models
• Visual and Vision-Language Pre-training
• AI Reasoning
• AI for Social Media
• Generative AI for Medical and Well-being

Media Security:

• Fake News/Media Detection
• Media Forensics and Anti-Forensics
• Adversarial Attack and Defense in AI-generated Media
• Deep Fakes/Misinformation/Disinformation
• Media/DNN Watermarking
• Reliability for Multimedia Applications and Systems
• The Security of Large AI Models
• Authenticity in Human-AI Generated Content
• Responsible AI for Multimedia

Are you interested in contributing a manuscript to this Research Topic?

New Trends in AI-Generated Media and Security