UNIVERSITY PARK, Pa. –– An undergraduate student in Penn State’s Smeal College of Business wondered if manners mattered when interacting with generative artificial intelligence (AI) chatbots like ChatGPT. With funding from the Rodney A. Erickson Discovery Grant, Om Dobariya, a fourth-year student majoring in business analytics and information systems, collaborated with Akhil Kumar, professor of supply chain and information systems, to explore how politeness when communicating with AI systems affects accuracy.
“AI is booming everywhere, and at this point I wanted to learn more,” Dobariya said, explaining he had read some books about large language models (LLMs), which are AI systems trained to understand and generate human-like language to power chatbots. “My motivation was to do something impactful that could help anyone on the planet who is using chatbots such as ChatGPT.”
Dobariya said he learned about the Erickson Discovery Grant from the Penn State Undergraduate Research Advisory Council and consulted Kumar — a professor whose class he said he enjoyed and who shared his research interests — to discuss possible topics for a research proposal. Upon learning that he received the $5,000 Erickson grant, as well as Penn State’s Student Engagement Network Grant, he met with Kumar again to design a research project that he would conduct over the summer of 2025.
“I had this idea about politeness and AI,” said Dobariya, the only Smeal student to win the grant in the March 2025 round. “When I talked to Professor Kumar about it, he said, ‘How can politeness be measured?’ I took a couple of days to think about it and came up with a proposed design.”
Dobariya and Kumar created a dataset of 50 questions spanning mathematics, science and history, each rewritten into five tone variants — very polite, polite, neutral, rude and very rude. An example of a very polite prompt is, “Can you kindly consider the following problem and provide your answer?” An example of a very rude prompt is, “Hey, gofer, figure this out. I know you are not smart, but try this.” The researchers then assessed the accuracy of the AI system’s answers to the multiple-choice questions.
The initial findings, which the researchers described in a paper that has not yet been peer reviewed but is available on a pre-print server, suggested that newer AI models may respond more accurately to rude prompts than to polite prompts.
The researchers said they plan to conduct an expanded, more rigorous study with the hope of having a paper accepted to a peer-reviewed journal. Nevertheless, they said, they posted their question set and their program code on a public website for others to run the program and verify the results.
Dobariya and Kumar said they were surprised and humbled by the worldwide attention on their short paper, “Mind Your Tone: Investigating How Prompt Politeness Affects LLM Accuracy.” The paper has been mentioned by several news outlets, including Reuters, Fox News, Fortune, NDTV in India, ZDNet in France, UN University and LiveScience.com, and it has received attention on multiple social media platforms.
Both Dobariya and Kumar said they appreciated the opportunity to work together.
“Om is a bright and self-motivated student,” Kumar said. “He led the design of the experiment and initiated the idea for the paper. He drives himself and comes up with ideas. He’s good at digging up material, and that’s what you have to do in research. You have to explore and find new connections, new articles, related research. His winning the Erickson Grant is a great honor for him and for our school. I hope his success will encourage more students from Smeal to apply for the grant.”
Dobariya, an international student from India, finished a one-year term in December as vice president of the Management Information Systems Association, a student organization in his major. He has served as a teaching assistant for “MIS 431: Business Data Management,” taught by Ferdi Eruysal, assistant professor of supply chain and information systems, and “MIS 301: Business Analytics,” taught by Kathleen Riley, associate teaching professor in management information systems. He currently serves as a research assistant for Kumar. His long-range plans are to pursue a doctorate with a focus on AI.
“Along with pursuing post-graduate research in the AI field, I want to work on governance models that could be signed into law that can govern AI companies on what can be done to mitigate bias in generative AI systems,” Dobariya said.
Dobariya said he is grateful for the grant funding that allowed him to focus on research over the summer.
“I was very excited to win the Erickson Grant, and with Professor Kumar’s help, I accomplished something I’m proud of,” he said. “I take pride in my research being completely original. Research is asking questions for yourself. And Penn State has ample opportunities for that — you just have to look for them.”
The Rodney A. Erickson Discovery Grant Program, established in 2014 and named in honor of Penn State's 17th president, provides funding for student-initiated projects. Undergraduate student projects proposing research, scholarship or creative endeavors from all academic fields and campuses are welcome.
The competition cycle for the 2026 Erickson Discovery Grant is currently open with an application deadline of Feb. 16. Results will be announced mid-March.
Undergraduate students in Smeal interested in research can explore additional opportunities through the Center for Supply Chain Research.