A new national study co-authored by New Mexico State University researcher Jin Ho Yun finds that students who rely on artificial intelligence systems such as ChatGPT or Gemini may develop more superficial and passive learning habits compared to those who conduct traditional online research. 

The study, published in PNAS Nexus, was led by Shiri Melumad of the University of Pennsylvania and Yun, an assistant professor of marketing in the NMSU College of Business. Their work examined how large language model (LLM) summaries shape learning outcomes and cognitive engagement. 

"We noticed a quiet but fundamental shift: AI systems don't just find information, they pre-package it," Yun said. "Traditional search forces you to pose queries, compare sources and synthesize. LLMs collapse that effort into a single summary, and our data show that convenience often trades off with depth of learning." 

"NMSU and the College of Business are at the forefront of exploring how AI will impact our society and how we learn," said Bryan Ashenbaum, dean of the College of Business. "This study by Dr. Yun and his colleague is a wonderful example of how our faculty engage with the opportunities and tradeoffs posed by AI and its growing adoption." 

Across seven preregistered online and lab experiments, the researchers found consistent results. Individuals who used AI-generated summaries reported feeling less engaged, retained less, and produced advice or written content that was less distinctive and less likely to be adopted by others. 

Yun said the team observed three clear patterns. Participants spent less time engaging with information when using AI, referenced fewer concrete facts, and produced ideas that were more templated and less original. "Less effort in, sparser and less original thinking out," he said. 

The study also found that even when the factual content was identical, participants felt they learned less from AI summaries. 

"It's not only what you read, but how you encounter it that shapes metacognition," Yun said. 

He also noted that AI can unintentionally make learning feel more passive by reducing the curiosity and exploration that occur during traditional online research. 

Yun emphasized that AI can play a productive role in education when used intentionally. He recommends assignments that require students to compare independent sources, reflect on how AI summaries shape their understanding, and create synthesis memos that integrate multiple inputs. He also encourages NMSU students across disciplines to strengthen their STEM and programming skills to better understand both AI's mechanisms and its broader philosophical implications. 

"We can design learning environments that treat AI as a collaborator rather than a crutch," Yun said. "The goal isn't to reject AI – it's to use it strategically and match the tool to the task." 

Yun added that NMSU's involvement in nationally visible research helps elevate the university's role in evidence-based AI pedagogy. "This project speaks to a core societal question – what AI does to human learning," he said. "Being part of work like this positions NMSU on the national stage and opens doors for future collaborations across education, technology and policy." 

The full article can be seen at https://newsroom.nmsu.edu/news/nmsu-researcher-co-leads-national-study-showing-ai-prompts-passive-learning/s/0f42a804-fc74-4047-a9df-88d43ae32512