`

When ChatGPT was launched in November 2022, it revolutionized the role of Artificial Intelligence (AI) in human lives. Developed by OpenAI, ChatGPT is a generative AI language model that creates customized text based on user input. Even with its cost-free features, it can assist users with tasks like writing, coding, research, and brainstorming. While its capability and accessibility are groundbreaking, the rise of generative AI also raises concerns about its impact on human intelligence, particularly in education.

In a 2024 survey that asked 490 university students from around the world about their reliance on ChatGPT, researchers found that 16.5% reported using ChatGPT for research and summarization, while 23.1% used ChatGPT to draft assignments. With so many students embedding AI into their learning process, is AI impacting student intelligence?

Anandi Nagarajan, Assistant Vice Provost for Pedagogy at NYU, suggested that its effects are open to mixed interpretations.

“In terms of the ways in which they're using it, I think there's a range,” said Nagarajan. “There might be students who kind of use it [to replace] their own learning, [some] students have started realizing it’s a pro and con.”

In her experience, she has observed that students seem to be ahead of faculty when it comes to adapting to AI, often finding creative ways to integrate AI into their learning process.

“I've heard other students use it in coding, in brainstorming ideas, or to get a short list of things to explore,” she said.

Students who tend to have better relationships with AI are the ones who use AI as a tool to enhance their learning productivity, rather than a placement of their critical thinking, where they stop questioning the validity of AI-generated responses.

“I think students who’ve developed that metacognitive reasoning to say ‘I should be more careful in what I’m getting out of ChatGPT’ are doing better,’” said Nagarajan. “They've realized that [they can’t] just take it and put it onto [their paper,] but need to have time to actually check whether this works or not.”

On the other hand, students who have a “one-shot deal” mindset, who simply insert their readings into ChatGPT and ask for a response to submit an assignment, tend to experience the more negative effects of AI.

Nagarajan said these students were not getting the best learning outcomes, as they rushed through assignments without fully engaging with the materials or evaluating AI-generated responses. Most class assignments are designed to help students grasp a solid understanding of the lessons, requiring them to engage with the material, explain their comprehension, offer their own insights, and even create their own contents. If students are using ChatGPT as a shortcut, students’ learning skills are actually being compromised.

“Students who might be thinking that anything that ChatGPT produces is perfect are the ones who are suffering learning loss,” she said. “They're not going to be learning because they might be trusting something that is not actually valid.”

When interpreting AI-generated responses, students should continue to question the trustworthiness of the content, similar to how they would when conducting research on their own.

“I think it’s very important to just develop critical thinking,” said Nagarajan. “I can’t assume that anything I read is a fact. I need to be asking myself the question ‘Does this make sense?’”

She recommended that when verifying the validity of AI-generated responses, students should be doing “data triangulation,” checking information across multiple sources to ensure accuracy.

“You want to make sure you're collecting data in multiple ways and checking that they're all saying the same thing,” said Nagarajan. “To ensure that your final interpretation makes sense and that it's justified by the evidence you've gathered”

Thus, it is an oversimplification to say that AI is making students “stupid.” Instead, the impact of AI on students’ intelligence should be examined through how students are integrating AI into their learning process. Those who experience negative effects are often over-reliant on AI and its responses, neglecting their own critical engagement with the material.

As AI is becoming more prevalent in almost all areas of life, Nagarajan advocated that faculty should not ban AI in classrooms as it will not adequately prepare students for the real world. Instead, faculty should reformulate their assignments to incorporate multiple levels of complexity.

“We are suggesting that you can’t use the simple assignments that we’ve had in the past that just have a question and the response can be generated by GenAI,” she said. “It has to be something that has a feedback loop built in, where you ask students to acknowledge, what are the sources that they’re utilizing to generate their research? and what does it really mean?”