A new US study reveals that relying on AI chatbots like ChatGPT or Gemini instead of traditional online research leads to superficial understanding and passive learning among students.

AI vs. Traditional Research: The Growing Debate


Artificial intelligence is transforming education, but not necessarily for the better. A team of US-based researchers has warned that using artificial intelligence instead of research could be making students and learners more superficial and passive in their pursuit of knowledge.

The study, published in PNAS Nexus, the journal of the US National Academy of Sciences, suggests that those who depend on large language models (LLMs) such as ChatGPT or Gemini develop a much shallower grasp of topics than those who engage in active online research.

According to lead authors Shiri Melumad of the University of Pennsylvania and Jin Ho Yun of New Mexico State University, AI-driven summaries may appear convenient, but they strip learners of the curiosity and engagement essential to deep understanding.

The Risk of Passive Learning


“Individuals who rely on LLM syntheses risk developing only a surface-level knowledge,” the researchers explained. Even when AI-generated summaries contain accurate information, they tend to make users less mentally active, reducing their capacity for analytical thinking and creativity.

The study involved seven online and laboratory experiments comparing participants who used AI tools with those who conducted standard online searches. The findings were consistent: AI users reported feeling less engaged and were less likely to produce original ideas or insights.

When participants relied on AI-generated summaries, their resulting advice or written content was “less distinctive and less likely to be adopted by others,” the researchers noted. This suggests that AI-based learning can limit not only comprehension but also originality.

Why Traditional Research Still Matters


Traditional online research encourages learners to explore, question, and connect diverse sources — all vital steps for building meaningful knowledge. In contrast, using artificial intelligence instead of research often skips the thinking process, offering neatly packaged answers that discourage deeper reflection.

Melumad and Yun emphasized that while AI systems make information retrieval faster, they may also “transform learning from an active to a passive experience,” leading to poorer learning outcomes over time.

The Future of Learning in the AI Era


As AI continues to integrate into classrooms and universities worldwide, experts warn against overreliance. Instead, they advocate a balanced approach — using AI as a complementary tool rather than a replacement for traditional inquiry.

The findings serve as a reminder that human curiosity, critical thinking, and exploration remain irreplaceable. In an era dominated by digital shortcuts, the call is clear: to learn deeply, we must keep learning actively — not passively through machines.

In the end, the message from researchers is simple yet powerful — using artificial intelligence instead of research makes us superficial and passive towards learning, but using both intelligently could redefine the way we understand knowledge itself.