An analysis of AI Decision under Risk: Prospect theory emerges in Large Language Models

TL;DR

This paper investigates whether large language models exhibit human-like risk judgment biases, specifically prospect theory, and finds that models do show framing effects influenced by scenario context, reflecting human heuristics and biases.

Contribution

It provides the first empirical evidence that large language models display prospect theory-like risk behaviors and highlights the importance of scenario framing in their decision-making processes.

Findings

Models exhibit prospect theory-like risk preferences.

Scenario framing significantly influences model risk judgments.

Military scenarios induce stronger framing effects than civilian ones.

Abstract

Judgment of risk is key to decision-making under uncertainty. As Daniel Kahneman and Amos Tversky famously discovered, humans do so in a distinctive way that departs from mathematical rationalism. Specifically, they demonstrated experimentally that humans accept more risk when they feel themselves at risk of losing something than when they might gain. I report the first tests of Kahneman and Tversky's landmark 'prospect theory' with Large Language Models, including today's state of the art chain-of-thought 'reasoners'. In common with humans, I find that prospect theory often anticipates how these models approach risky decisions across a range of scenarios. I also demonstrate that context is key to explaining much of the variance in risk appetite. The 'frame' through which risk is apprehended appears to be embedded within the language of the scenarios tackled by the models. Specifically, I find that military scenarios generate far larger 'framing effects' than do civilian settings, ceteris paribus. My research suggests, therefore, that language models the world, capturing our human heuristics and biases. But also that these biases are uneven - the idea of a 'frame' is richer than simple gains and losses. Wittgenstein's notion of 'language games' explains the contingent, localised biases activated by these scenarios. Finally, I use my findings to reframe the ongoing debate about reasoning and memorisation in LLMs.