VideoVeritas: AI-Generated Video Detection via Perception Pretext Reinforcement Learning

TL;DR

VideoVeritas introduces a novel framework combining perception and reasoning to improve detection of AI-generated videos, utilizing reinforcement learning and a new dataset to enhance robustness and balance in performance.

Contribution

The paper proposes a new perception pretext reinforcement learning approach and a high-quality dataset, advancing AI-generated video detection beyond existing methods.

Findings

VideoVeritas outperforms existing detection methods across multiple benchmarks.

The framework achieves balanced reasoning and perception capabilities.

The MintVid dataset provides a valuable resource for future research.

Abstract

The growing capability of video generation poses escalating security risks, making reliable detection increasingly essential. In this paper, we introduce VideoVeritas, a framework that integrates fine-grained perception and fact-based reasoning. We observe that while current multi-modal large language models (MLLMs) exhibit strong reasoning capacity, their granular perception ability remains limited. To mitigate this, we introduce Joint Preference Alignment and Perception Pretext Reinforcement Learning (PPRL). Specifically, rather than directly optimizing for detection task, we adopt general spatiotemporal grounding and self-supervised object counting in the RL stage, enhancing detection performance with simple perception pretext tasks. To facilitate robust evaluation, we further introduce MintVid, a light yet high-quality dataset containing 3K videos from 9 state-of-the-art generators, along with a real-world collected subset that has factual errors in content. Experimental results demonstrate that existing methods tend to bias towards either superficial reasoning or mechanical analysis, while VideoVeritas achieves more balanced performance across diverse benchmarks.