COMIC: Agentic Sketch Comedy Generation

TL;DR

COMIC is an automated AI system that generates short comedic videos by simulating production roles, using large language model critics aligned with viewer preferences, achieving high-quality results comparable to professional sketches.

Contribution

The paper introduces a novel AI framework for comedy video creation that combines agent-based role simulation with LLM critics trained on viewer preferences.

Findings

Produces comedy videos approaching professional quality

Demonstrates state-of-the-art video generation performance

Aligns humor evaluation with viewer preferences

Abstract

We propose a fully automated AI system that produces short comedic videos similar to sketch shows such as Saturday Night Live. Starting with character references, the system employs a population of agents loosely based on real production studio roles, structured to optimize the quality and diversity of ideas and outputs through iterative competition, evaluation, and improvement. A key contribution is the introduction of LLM critics aligned with real viewer preferences through the analysis of a corpus of comedy videos on YouTube to automatically evaluate humor. Our experiments show that our framework produces results approaching the quality of professionally produced sketches while demonstrating state-of-the-art performance in video generation.