SocialDF: Benchmark Dataset and Detection Model for Mitigating Harmful Deepfake Content on Social Media Platforms

TL;DR

This paper introduces SocialDF, a comprehensive dataset of real-world deepfakes on social media, and a novel multi-factor detection model leveraging large language models to improve deepfake identification.

Contribution

The paper presents a new benchmark dataset, SocialDF, and a multi-modal detection approach using LLMs, advancing deepfake detection capabilities on social media platforms.

Findings

SocialDF covers diverse real-world deepfakes from online sources.

The LLM-based detection model effectively combines multiple verification factors.

Results demonstrate improved accuracy over existing detection methods.

Abstract

The rapid advancement of deep generative models has significantly improved the realism of synthetic media, presenting both opportunities and security challenges. While deepfake technology has valuable applications in entertainment and accessibility, it has emerged as a potent vector for misinformation campaigns, particularly on social media. Existing detection frameworks struggle to distinguish between benign and adversarially generated deepfakes engineered to manipulate public perception. To address this challenge, we introduce SocialDF, a curated dataset reflecting real-world deepfake challenges on social media platforms. This dataset encompasses high-fidelity deepfakes sourced from various online ecosystems, ensuring broad coverage of manipulative techniques. We propose a novel LLM-based multi-factor detection approach that combines facial recognition, automated speech transcription, and a multi-agent LLM pipeline to cross-verify audio-visual cues. Our methodology emphasizes robust, multi-modal verification techniques that incorporate linguistic, behavioral, and contextual analysis to effectively discern synthetic media from authentic content.