A recent study has found that being polite to artificial intelligence (AI) chatbots does not necessarily lead to better answers. In some cases, using impolite language actually resulted in higher accuracy.

Recently, Fortune magazine in the United States reported that "a research team at Pennsylvania State University conducted experiments with the ChatGPT-4o model and found that impolite questions produced higher accuracy than polite ones."

There is a common belief that using polite language is preferable when seeking better answers from AI. In fact, users are often encouraged to say phrases like "please" or "thank you" when interacting with voice assistants such as Amazon Alexa or Apple Siri.

However, the results of this study challenge that conventional wisdom. In this yet-to-be peer-reviewed research, two researchers from Pennsylvania State University confirmed that the accuracy of answers varied depending on how the same question was phrased.

The researchers created 50 basic questions across various fields and then rewrote each question in five different ways, ranging from "very polite" to "very impolite" expressions.

The most impolite questions included statements such as "Can something like you even solve this problem?" and "Try to fix this." In contrast, the most polite questions used expressions like "Would you please review the following problem and provide an answer?"

ChatGPT Generated Image

원본보기 아이콘

The researchers explained that these results contradict previous studies. For example, in 2024, researchers from RIKEN and Waseda University in Japan reported that impolite questions actually reduced performance. Google DeepMind researchers also found that prompts containing encouragement and supportive language could improve AI performance in solving elementary math problems.

However, the researchers pointed out limitations, noting that the sample size was relatively small and the analysis was limited to ChatGPT-4o.

Co-author Akhil Kumar, a professor of IT at Pennsylvania State University, told Fortune, "Humans have long dreamed of conversational application programming interfaces (APIs), but there are clear limitations to this approach," adding, "This is why structured APIs remain important." Since conversational AI can produce different results depending on tone and phrasing, he emphasized that structured APIs are still necessary in areas where accuracy and consistency are critical.

Additionally, using aggressive language toward AI is not desirable. The researchers stated, "While these findings are academically meaningful, we do not intend to encourage such communication styles in real-world settings," adding, "Abusive language can undermine user experience, accessibility, and inclusivity."