What Does the Bot Say? Opportunities and Risks of Large Language Models in Social Media Bot Detection
Instruction-tuned LLMs with a mixture-of-experts framework outperform bot detectors by 9.1%, but LLM-guided manipulation reduces their performance by up to 29.6%.
Published 202427 citationsPaper ↗
Only vote on papers you've read. Sign in with GitHub to vote.
A rigorous study that establishes LLMs as the new bot-detection frontier with striking 9.1% gains and 29.6% evasion drops, though it leaves calibration collapse, inference cost, and the true attack surface of cheap LLM-guided manipulation unresolved.
Abstract
Social media bot detection has always been an arms race between advancements in machine learning bot detectors and adversarial bot strategies to evade detection.In this work, we bring the arms race to the next level by investigating the opportunities and risks of state-of-the-art large language models (LLMs) in social bot detection.To investigate the opportunities, we design novel LLM-based bot detectors by proposing a mixture-of-heterogeneous-experts framework to divide and conquer diverse user information modalities.To illuminate the risks, we explore the possibility of LLM-guided manipulation of user textual and structured information to evade detection.Extensive experiments with three LLMs on two datasets demonstrate that instruction tuning on merely 1,000 annotated examples produces specialized LLMs that outperform state-of-the-art bot detection baselines by up to 9.1% on both datasets.On the other hand, LLM-guided manipulation strategies could significantly bring down the performance of existing bot detectors by up to 29.6% and harm the calibration and reliability of bot detection systems.Ultimately, this works identifies LLMs as the new frontier of social bot detection research.1