Revisiting Early Detection of Sexual Predators via Turn-level Optimization

2025-03-09

AI Summary

Early intervention against online grooming faces a critical challenge: conventional chat-level risk assessment lacks the temporal precision required to identify the optimal intervention point. To address this, we propose a turn-level dynamic modeling framework grounded in the Luring Communication Theory (LCT), enabling fine-grained turn-level risk annotation and overcoming the limitations of static, coarse-grained evaluation. We further introduce Speed-Controlled Reinforcement Learning (SCoRL), a novel RL-based method incorporating a speed–accuracy trade-off reward function and turn-level evaluation metrics to enable real-time tracking of grooming progression and precise identification of the optimal intervention turn. Evaluated on authentic grooming dialogues, our approach triggers intervention an average of 1.8 turns earlier than baselines and achieves a 23.6% improvement in F1@early, significantly enhancing both the timeliness and interpretability of proactive protection.

Technology Category

• Natural Language Processing: Safety and Robustness
• Machine Learning: Online Learning & Bandits
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Online grooming is a severe social threat where sexual predators gradually entrap child victims with subtle and gradual manipulation. Therefore, timely intervention for online grooming is critical for proactive protection. However, previous methods fail to determine the optimal intervention points (i.e., jump to conclusions) as they rely on chat-level risk labels by causing weak supervision of risky utterances. For timely detection, we propose speed control reinforcement learning (SCoRL) (The code and supplementary materials are available at https://github.com/jinmyeongAN/SCoRL), incorporating a practical strategy derived from luring communication theory (LCT). To capture the predator's turn-level entrapment, we use a turn-level risk label based on the LCT. Then, we design a novel speed control reward function that balances the trade-off between speed and accuracy based on turn-level risk label; thus, SCoRL can identify the optimal intervention moment. In addition, we introduce a turn-level metric for precise evaluation, identifying limitations in previously used chat-level metrics. Experimental results show that SCoRL effectively preempted online grooming, offering a more proactive and timely solution. Further analysis reveals that our method enhances performance while intuitively identifying optimal early intervention points.

Problem

• Detect optimal intervention points in online grooming
• Balance speed and accuracy in predator detection
• Evaluate grooming detection with turn-level metrics

Innovation

• Speed control reinforcement learning for detection
• Turn-level risk labels based on LCT
• Novel speed control reward function