# Bot Wars Evolved: Orchestrating Competing LLMs in a Counterstrike Against Phone Scams

arXiv: https://arxiv.org/abs/2503.07036 · HTML: https://mlsanity.com/paper/2503.07036/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 13 references that are in the map.
- First posted on arXiv 10 Mar 2025.

## Builds on (most cited first, 4 of 4)

- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [Challenges in Building Intelligent Open-domain Dialog Systems](https://mlsanity.com/paper/1905.05709/) (2019)
- [A Deep Reinforcement Learning Chatbot](https://mlsanity.com/paper/1709.02349/) (2017)
- [Automatic Scam-Baiting Using ChatGPT](https://mlsanity.com/paper/2309.01586/) (2023)

## Built on by (most cited first, 5 of 5)

- [A Survey of Agentic AI and Cybersecurity: Challenges, Opportunities and Use-case Prototypes](https://mlsanity.com/paper/2601.05293/) (2026)
- [Send to which account? Evaluation of an LLM-based Scambaiting System](https://mlsanity.com/paper/2509.08493/) (2025)
- [DIALEVAL: Automated Type-Theoretic Evaluation of LLM Instruction Following](https://mlsanity.com/paper/2603.03321/) (2026)
- [ORACLE: Anticipating Scams from Partial Trajectories in Streaming App Usage](https://mlsanity.com/paper/2605.16363/) (2026)
- [An Expanded Synthetic Conversation Dataset for Multi-Turn Smishing Detection](https://mlsanity.com/paper/2606.06879/) (2026)

## Abstract (arXiv)

We present "Bot Wars," a framework using Large Language Models (LLMs) scam-baiters to counter phone scams through simulated adversarial dialogues. Our key contribution is a formal foundation for strategy emergence through chain-of-thought reasoning without explicit optimization. Through a novel two-layer prompt architecture, our framework enables LLMs to craft demographically authentic victim personas while maintaining strategic coherence. We evaluate our approach using a dataset of 3,200 scam dialogues validated against 179 hours of human scam-baiting interactions, demonstrating its effectiveness in capturing complex adversarial dynamics. Our systematic evaluation through cognitive, quantitative, and content-specific metrics shows that GPT-4 excels in dialogue naturalness and persona authenticity, while Deepseek demonstrates superior engagement sustainability.