Discovering Coordinated Processes From Social Online Networks

2025-06-15

AI Summary

Detecting coordinated disinformation campaigns involving both AI and human agents in social networks remains challenging, as conventional diffusion models fail to capture complex coordination structures—such as multi-path propagation, branching, and loops. Method: This work introduces a process-mining–based approach to identify coordinated behavioral patterns. Specifically, it adapts process discovery algorithms (e.g., Inductive Miner) to social network metadata—particularly post timestamps—to automatically uncover latent control-flow structures; these are then integrated with temporal graph modeling to enable interpretable behavioral provenance tracing. Contribution/Results: Evaluated on real-world Twitter/X event data, the method significantly improves discrimination accuracy between AI-driven and human-initiated dissemination patterns. It establishes a novel paradigm for malicious coordination detection that jointly achieves high precision and model interpretability—addressing critical gaps in transparency and accountability within automated disinformation analysis.

Technology Category

• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust
• Humans and AI: Planning and Decision Support for Human-Machine Teams
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Web Mining and Content Analysis: Content-based information diffusion
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

The rapid growth of social media presents a unique opportunity to study coordinated agent behavior in an unfiltered environment. Online processes often exhibit complex structures that reflect the nature of the user behavior, whether it is authentic and genuine, or part of a coordinated effort by malicious agents to spread misinformation and disinformation. Detection of AI-generated content can be extremely challenging due to the high quality of large language model-generated text. Therefore, approaches that use metadata like post timings are required to effectively detect coordinated AI-driven campaigns. Existing work that models the spread of information online is limited in its ability to represent different control flows that occur within the network in practice. Process mining offers techniques for the discovery of process models with different routing constructs and are yet to be applied to social networks. We propose to leverage process mining methods for the discovery of AI and human agent behavior within social networks. Applying process mining techniques to real-world Twitter (now X) event data, we demonstrate how the structural and behavioral properties of discovered process models can reveal coordinated AI and human behaviors online.

Problem

• Detect coordinated AI and human behavior in social networks
• Identify misinformation spread using process mining techniques
• Analyze structural properties to reveal online coordination

Innovation

• Leverage process mining for behavior discovery
• Analyze metadata to detect coordinated campaigns
• Model structural properties of online processes