Measuring Human and AI Values Based on Generative Psychometrics with Large Language Models

TL;DR

This paper introduces Generative Psychometrics for Values (GPV), a novel LLM-based method for measuring human and AI values through text analysis, demonstrating its effectiveness and potential for value-aligned AI development.

Contribution

The paper presents a new psychometric paradigm for measuring values using LLM outputs, extending traditional psychometrics to AI and providing a comparative analysis of measurement methods.

Findings

GPV is stable and valid for human blog analysis

GPV outperforms prior psychological tools in value measurement

LMM value measurement can predict safety and value alignment

Abstract

Human values and their measurement are long-standing interdisciplinary inquiry. Recent advances in AI have sparked renewed interest in this area, with large language models (LLMs) emerging as both tools and subjects of value measurement. This work introduces Generative Psychometrics for Values (GPV), an LLM-based, data-driven value measurement paradigm, theoretically grounded in text-revealed selective perceptions. The core idea is to dynamically parse unstructured texts into perceptions akin to static stimuli in traditional psychometrics, measure the value orientations they reveal, and aggregate the results. Applying GPV to human-authored blogs, we demonstrate its stability, validity, and superiority over prior psychological tools. Then, extending GPV to LLM value measurement, we advance the current art with 1) a psychometric methodology that measures LLM values based on their scalable and free-form outputs, enabling context-specific measurement; 2) a comparative analysis of measurement paradigms, indicating response biases of prior methods; and 3) an attempt to bridge LLM values and their safety, revealing the predictive power of different value systems and the impacts of various values on LLM safety. Through interdisciplinary efforts, we aim to leverage AI for next-generation psychometrics and psychometrics for value-aligned AI.