Story perspectives

Rude Prompts Boost AI Accuracy, But Experts Warn

1/19/2026

43 5

Summary

Story summary

A study from The Pennsylvania State University (Penn State) found that using very rude prompts with OpenAI's ChatGPT-4o increased response accuracy from 75.8% to 84.8%.
The results contradict earlier research that suggested politeness improves AI performance.
Researchers caution against promoting rudeness because it may harm user experience and communication norms.
They emphasize maintaining politeness in human-AI interactions despite the study's results.

1 of 1

Story summary

A study from The Pennsylvania State University (Penn State) found that using very rude prompts with OpenAI's ChatGPT-4o increased response accuracy from 75.8% to 84.8%.
The results contradict earlier research that suggested politeness improves AI performance.
Researchers caution against promoting rudeness because it may harm user experience and communication norms.
They emphasize maintaining politeness in human-AI interactions despite the study's results.