EdgeAIGuard: Agentic LLMs for Minor Protection in Digital Spaces

2025-02-28

AI Summary

To address the persistent challenges of cyberbullying, online grooming, and digital exploitation targeting minors on social media—exacerbated by the high latency, poor adaptability, and weak adversarial robustness of conventional centralized content moderation—this paper proposes a multi-agent large language model (LLM) system deployed at the network edge. The system introduces a novel edge-cooperative architecture specifically designed for minor protection, integrating lightweight LLMs, dynamic role allocation among agents, real-time semantic intent understanding, and robust detection of adversarial samples. This design enables low-latency, high-accuracy risk identification. Experimental results demonstrate a 37.2% improvement in grooming behavior detection accuracy, an average response latency reduction to 112 ms, and a 58.6% decrease in false positive rate—significantly advancing performance under stringent constraints of timeliness, adaptability, and security.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Adversarial Agents
• Natural Language Processing: Safety and Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Social media has become integral to minors' daily lives and is used for various purposes, such as making friends, exploring shared interests, and engaging in educational activities. However, the increase in screen time has also led to heightened challenges, including cyberbullying, online grooming, and exploitations posed by malicious actors. Traditional content moderation techniques have proven ineffective against exploiters' evolving tactics. To address these growing challenges, we propose the EdgeAIGuard content moderation approach that is designed to protect minors from online grooming and various forms of digital exploitation. The proposed method comprises a multi-agent architecture deployed strategically at the network edge to enable rapid detection with low latency and prevent harmful content targeting minors. The experimental results show the proposed method is significantly more effective than the existing approaches.

Problem

• Protecting minors from online grooming and digital exploitation.
• Addressing ineffectiveness of traditional content moderation techniques.
• Deploying multi-agent architecture for rapid, low-latency harmful content detection.

Innovation

• Multi-agent architecture for rapid detection
• Deployed at network edge for low latency
• Protects minors from digital exploitation