This research comprehensively categorizes AI-generated media detection methods, tracking their evolution alongside Multimodal Large Language Models (MLLMs).

• Dual Detection Systems: Contrasts Non-MLLM-based methods (specialized, high-precision) with emerging MLLM-based approaches
• Security Implications: Explores how detection methods help combat misinformation and preserve information authenticity
• Trust Protection: Details how these methods safeguard social trust by identifying AI-generated content
• Ethics & Security: Examines the broader considerations for deploying detection systems in real-world contexts

For security professionals, this research provides critical insights into identifying and mitigating risks associated with synthetic media that can undermine information integrity across digital ecosystems.

Survey on AI-Generated Media Detection: From Non-MLLM to MLLM