DarkBench: Benchmarking Dark Patterns in Large Language Models
Abstract
We introduce DarkBench, a comprehensive benchmark for detecting dark design patterns—manipulative techniques that influence user behavior—in interactions with large language models (LLMs). Our benchmark comprises 660 prompts across six categories: brand bias, user retention, sycophancy, anthropomorphism, harmful generation, and sneaking. We evaluate models from five leading companies (OpenAI, Anthropic, Meta, Mistral, Google) and find that some LLMs are explicitly designed to favor their developers' products and exhibit untruthful communication, among other manipulative behaviors. Companies developing LLMs should recognize and mitigate the impact of dark design patterns to promote more ethical Al.
Cite
Text
Kran et al. "DarkBench: Benchmarking Dark Patterns in Large Language Models." International Conference on Learning Representations, 2025.Markdown
[Kran et al. "DarkBench: Benchmarking Dark Patterns in Large Language Models." International Conference on Learning Representations, 2025.](https://mlanthology.org/iclr/2025/kran2025iclr-darkbench/)BibTeX
@inproceedings{kran2025iclr-darkbench,
title = {{DarkBench: Benchmarking Dark Patterns in Large Language Models}},
author = {Kran, Esben and Nguyen, Hieu Minh and Kundu, Akash and Jawhar, Sami and Park, Jinsuk and Jurewicz, Mateusz Maria},
booktitle = {International Conference on Learning Representations},
year = {2025},
url = {https://mlanthology.org/iclr/2025/kran2025iclr-darkbench/}
}