Anthropomimetic Uncertainty: What Verbalized Uncertainty in Language Models is Missing

TL;DR

This paper explores how language models can better communicate their uncertainty through human-like linguistic cues, addressing overconfidence issues and improving trustworthiness in human-AI interactions.

Contribution

It introduces the concept of anthropomimetic uncertainty, emphasizing the importance of mimicking human uncertainty communication in NLP models and provides a comprehensive analysis of related biases.

Findings

Identifies biases affecting verbalized uncertainty in NLP

Highlights the gap between human and machine uncertainty communication

Proposes future research directions for anthropomimetic uncertainty

Abstract

Human users increasingly communicate with large language models (LLMs), but LLMs suffer from frequent overconfidence in their output, even when its accuracy is questionable, which undermines their trustworthiness and perceived legitimacy. Therefore, there is a need for language models to signal their confidence in order to reap the benefits of human-machine collaboration and mitigate potential harms. Verbalized uncertainty is the expression of confidence with linguistic means, an approach that integrates perfectly into language-based interfaces. Most recent research in natural language processing (NLP) overlooks the nuances surrounding human uncertainty communication and the biases that influence the communication of and with machines. We argue for anthropomimetic uncertainty, the principle that intuitive and trustworthy uncertainty communication requires a degree of imitation of human linguistic behaviors. We present a thorough overview of the research in human uncertainty communication, survey ongoing research in NLP, and perform additional analyses to demonstrate so-far underexplored biases in verbalized uncertainty. We conclude by pointing out unique factors in human-machine uncertainty and outlining future research directions towards implementing anthropomimetic uncertainty.