Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces StealthRL, a reinforcement learning framework that generates adversarial paraphrases to evade AI-text detectors. StealthRL trains a Qwen3-4B paraphrase policy using Group Relative Policy Optimization (GRPO) with LoRA adapters, optimizing for both detector evasion and semantic similarity against an ensemble of detectors. Experiments across six attack settings and three detector families demonstrate StealthRL's ability to achieve near-zero detection rates (0.001 TPR@1%FPR) and high attack success rates (99.9%), even transferring to unseen detector families.
Current AI-text detectors are surprisingly brittle: a novel RL-based paraphrasing attack can reduce their detection ability to near-zero while preserving meaning, even against detectors they weren't trained on.
AI-text detectors face a critical robustness challenge: adversarial paraphrasing attacks that preserve semantics while evading detection. We introduce StealthRL, a reinforcement learning framework that stress-tests detector robustness under realistic adversarial conditions. StealthRL trains a paraphrase policy against a multi-detector ensemble using Group Relative Policy Optimization (GRPO) with LoRA adapters on Qwen3-4B, optimizing a composite reward that balances detector evasion with semantic preservation. We evaluate six attack settings (M0-M5) against three detector families (RoBERTa, FastDetectGPT, and Binoculars) at the security-relevant 1% false positive rate operating point. StealthRL achieves near-zero detection (0.001 mean TPR@1%FPR), reduces mean AUROC from 0.74 to 0.27, and attains a 99.9% attack success rate. Critically, attacks transfer to a held-out detector family not seen during training, revealing shared architectural vulnerabilities rather than detector-specific brittleness. We additionally conduct LLM-based quality evaluation via Likert scoring, analyze detector score distributions to explain why evasion succeeds, and provide per-detector AUROC with bootstrap confidence intervals. Our results expose significant robustness gaps in current AI-text detection and establish StealthRL as a principled adversarial evaluation protocol. Code and evaluation pipeline are publicly available at https://github.com/suraj-ranganath/StealthRL.