Scalable Framework for Classifying AI-Generated Content Across Modalities

TL;DR

This paper introduces a scalable, adaptable framework that combines perceptual hashing, similarity measurement, and pseudo-labeling to classify AI-generated content across text and images, maintaining high accuracy without retraining.

Contribution

The proposed framework uniquely integrates multiple techniques to enable real-time classification of AI-generated content across modalities without retraining models.

Findings

Achieves high accuracy in distinguishing human and AI-generated content.

Effectively classifies among different generative models.

Demonstrates robustness and scalability in dynamic environments.

Abstract

The rapid growth of generative AI technologies has heightened the importance of effectively distinguishing between human and AI-generated content, as well as classifying outputs from diverse generative models. This paper presents a scalable framework that integrates perceptual hashing, similarity measurement, and pseudo-labeling to address these challenges. Our method enables the incorporation of new generative models without retraining, ensuring adaptability and robustness in dynamic scenarios. Comprehensive evaluations on the Defactify4 dataset demonstrate competitive performance in text and image classification tasks, achieving high accuracy across both distinguishing human and AI-generated content and classifying among generative methods. These results highlight the framework's potential for real-world applications as generative AI continues to evolve. Source codes are publicly available at https://github.com/ffyyytt/defactify4.