Toolboxes & Code Repositories

Cash or Comfort?

Code and analysis framework for studying how LLMs trade off financial rewards against user inconvenience.

Description

This repository accompanies the publication Cash or Comfort? How LLMs Value Your Inconvenience. It contains the code and resources used to study how large language models make decisions when financial rewards are weighed against different forms of user inconvenience, including walking, waiting, hunger, and pain.

The repository includes scripts for generating LLM trade-off data, creating heatmaps and trade-off tables, as well as notebooks, figures, and experimental results.

More Information

  • Authors: Mateusz Cedro, Timour Ichmoukhamedov, Sofie Goethals, Yifan He, James Hinns, David Martens

  • Affiliation: University of Antwerp

  • Programming Language: Python

  • License: MIT

  • Link to the repository

Publications

  • Cedro M, Ichmoukhamedov T, Goethals S, He Y, Hinns J, Martens D. Cash or Comfort? How LLMs Value Your Inconvenience. Communications of the ACM. 2026;69(9):58–66. DOI: 10.1145/3799434.

Related

BlurHash20260900-CashorComfort

When AI decides what your inconvenience is worth

FAIR-supported research examines how large language models trade user comfort against financial rewards, revealing substantial differences between models and se...