VidGuard-R1: AI-Generated Video Detection and Explanation via Reasoning MLLMs and RL

2025-10-02

AI Summary

To address the growing misinformation and reputational risks posed by AI-generated videos, this paper proposes the first multimodal framework that jointly performs detection and interpretable attribution. Methodologically, it introduces a novel integration of fine-tuned Qwen-VL—a large multimodal foundation model—with Group Relative Policy Optimization (GRPO), coupled with a dual reward modeling scheme: one reward model captures temporal artifacts, while the other quantifies generation complexity, enabling end-to-end detection and attribution in a unified architecture. Trained on a dataset of 140K real and synthetic videos, the framework achieves state-of-the-art zero-shot generalization performance, with >95% accuracy. Case studies demonstrate its capability to generate precise, semantically coherent attribution explanations. This work establishes a trustworthy, high-accuracy, and highly transparent paradigm for governing AI-generated video content.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Philosophy and Ethics of AI: Accountability, Interpretability & Explainability

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

With the rapid advancement of AI-generated videos, there is an urgent need for effective detection tools to mitigate societal risks such as misinformation and reputational harm. In addition to accurate classification, it is essential that detection models provide interpretable explanations to ensure transparency for regulators and end users. To address these challenges, we introduce VidGuard-R1, the first video authenticity detector that fine-tunes a multi-modal large language model (MLLM) using group relative policy optimization (GRPO). Our model delivers both highly accurate judgments and insightful reasoning. We curate a challenging dataset of 140k real and AI-generated videos produced by state-of-the-art generation models, carefully designing the generation process to maximize discrimination difficulty. We then fine-tune Qwen-VL using GRPO with two specialized reward models that target temporal artifacts and generation complexity. Extensive experiments demonstrate that VidGuard-R1 achieves state-of-the-art zero-shot performance on existing benchmarks, with additional training pushing accuracy above 95%. Case studies further show that VidGuard-R1 produces precise and interpretable rationales behind its predictions. The code is publicly available at https://VidGuard-R1.github.io.

Problem

• Detecting AI-generated videos to combat misinformation risks
• Providing interpretable explanations for video authenticity decisions
• Addressing temporal artifacts and generation complexity in detection

Innovation

• Fine-tunes MLLM using group relative policy optimization
• Employs reward models for temporal artifacts and complexity
• Achieves high accuracy with interpretable prediction rationales