Multi-Agent Comedy Club: Investigating Community Discussion Effects on LLM Humor Generation

TL;DR

This study investigates how community discussion influences humor generation in large language models, showing that incorporating social memory through discussion enhances humor quality and audience engagement in a controlled multi-agent environment.

Contribution

It introduces a novel multi-agent framework that integrates community discussion as social memory to improve LLM humor generation, a previously underexplored aspect.

Findings

Discussion condition outperforms baseline in humor quality (75.6% wins).

Increases in Craft/Clarity and Social Response scores.

Occasional rise in aggressive humor.

Abstract

Prior work has explored multi-turn interaction and feedback for LLM writing, but evaluations still largely center on prompts and localized feedback, leaving persistent public reception in online communities underexamined. We test whether broadcast community discussion improves stand-up comedy writing in a controlled multi-agent sandbox: in the discussion condition, critic and audience threads are recorded, filtered, stored as social memory, and later retrieved to condition subsequent generations, whereas the baseline omits discussion. Across 50 rounds (250 paired monologues) judged by five expert annotators using A/B preference and a 15-item rubric, discussion wins 75.6% of instances and improves Craft/Clarity ({\Delta} = 0.440) and Social Response ({\Delta} = 0.422), with occasional increases in aggressive humor.