CulturalTeaming: AI-Assisted Interactive Red-Teaming for Challenging LLMs' (Lack of) Multicultural Knowledge

TL;DR

CulturalTeaming is an interactive AI-assisted system that enhances the creation of multicultural evaluation datasets for LLMs by combining human expertise and AI support, revealing significant gaps in LLMs' multicultural knowledge.

Contribution

The paper introduces CulturalTeaming, a novel human-AI collaborative system for generating challenging multicultural evaluation datasets for LLMs, improving annotation quality and diversity.

Findings

CulturalTeaming's AI assistance helps create more challenging cultural questions.

Participants felt more creative and confident with increased AI support.

Modern LLMs' accuracy on CULTURALBENCH-V0.1 ranges from 37.7% to 72.2%, indicating a knowledge gap.

Abstract

Frontier large language models (LLMs) are developed by researchers and practitioners with skewed cultural backgrounds and on datasets with skewed sources. However, LLMs' (lack of) multicultural knowledge cannot be effectively assessed with current methods for developing benchmarks. Existing multicultural evaluations primarily rely on expensive and restricted human annotations or potentially outdated internet resources. Thus, they struggle to capture the intricacy, dynamics, and diversity of cultural norms. LLM-generated benchmarks are promising, yet risk propagating the same biases they are meant to measure. To synergize the creativity and expert cultural knowledge of human annotators and the scalability and standardizability of LLM-based automation, we introduce CulturalTeaming, an interactive red-teaming system that leverages human-AI collaboration to build truly challenging evaluation dataset for assessing the multicultural knowledge of LLMs, while improving annotators' capabilities and experiences. Our study reveals that CulturalTeaming's various modes of AI assistance support annotators in creating cultural questions, that modern LLMs fail at, in a gamified manner. Importantly, the increased level of AI assistance (e.g., LLM-generated revision hints) empowers users to create more difficult questions with enhanced perceived creativity of themselves, shedding light on the promises of involving heavier AI assistance in modern evaluation dataset creation procedures. Through a series of 1-hour workshop sessions, we gather CULTURALBENCH-V0.1, a compact yet high-quality evaluation dataset with users' red-teaming attempts, that different families of modern LLMs perform with accuracy ranging from 37.7% to 72.2%, revealing a notable gap in LLMs' multicultural proficiency.