How, why, and with what cognitive consequences do students use AI?

Triple Five

Nipun Ranchhod, NAVADIA, Pace University Tewan, POTHIBUTT, Pace University Kanokkan, KANJANAPIBAN, New York University

Training / education / pedagogy

To have the real understanding of how AI is being used we have initiated student surveys for analysis. We have prepared a very intensive survey questionnaire which completely relates to student life. We have initiated this survey drive in various universities for better perspective. The survey has been initiated in Pace University which is very well known for its profession-focused approach, New York University, which is private large-scale university known for its research oriented methods for teaching and learning, Columbia University to get insights about one of the best Ivy League institutions, colleges from City University of New York which is a premier public university known for high quality education like Baruch college, Hunter college, City college of NY, Queens college. As per the responses we received, students have share concerns about accuracy of the responses due to hallucinations and also about the security as that's always a major factor whenever anything gets to market and is in use.

The study performed by researchers in 2025 shows a comparative analysis to understand how the brain is affected by the use of AI. They decided to have 3 groups of student, the first group, ‘A ’ was asked to write an essay without using any external help on their own, the second group, ‘B’ was asked to use search engines to take help for the same task, and the third group ‘C’, was asked to use LLMs for writing the essay. All groups were in the same surroundings, and the ecosystem had only the constraints and instructions imposed by resources. They identified that the writing is linked to cognitive engagement and brain connectivity. This showed that the group using LLMs had weaker brain connectivity and cognitive acceleration than group A.

Everyone is trying to bound the use of AI by restrictions or want it to be used with no concern about the threats and risks it may pose to every individual physically (by sharing personal or any other details) or mentally (cognitive impacts). We don’t want to restrict the use and police AI but we would like to explore the scenarios in depth and partner with AI. The ecosystem where students partner with AI and create an environment for student and AI literacy.

How, why, and with what cognitive consequences do students use AI? What is the situation or context you are addressing? In present we all use AI for our day to day tasks or even to get answers about simple things or even to check what to reply to texts. Students are one of the major users of AI, it is used for learning/understanding concepts, assistance for projects or assignments and many more reasons. This reliance on students at the learning phase also continues after they complete their student life. This reliance in real world makes them cognitively slow along with getting critical thinking ability slower. One of the reasons for this hit on cognitive engagement and critical thinking is acceptance of response without confirming the facts or sources which pushes AI to hallucination and on other hand students don't challenge their thoughts even if the responses from AI are wrong [1]. The study performed by researchers in 2025 shows a comparative analysis to understand how the brain is affected by the use of AI. They decided to have 3 groups of student, the first group, ‘A ’ was asked to write an essay without using any external help on their own, the second group, ‘B’ was asked to use search engines to take help for the same task, and the third group ‘C’, was asked to use LLMs for writing the essay. All groups were in the same surroundings, and the ecosystem had only the constraints and instructions imposed by resources. They identified that the writing is linked to cognitive engagement and brain connectivity. This showed that the group using LLMs had weaker brain connectivity and conative acceleration than group A [2]. From this study, we got an understanding that with the impact on the brain, students are under the influence of cognitive offloading as they are getting habituated to outsourcing the tasks they find difficult to the LLMs for faster and easier execution and completion however, this is causing another problem of cognitive atrophy under which the students are losing ability of critical thinking and that leads to another problem of problem solving. We as students can totally relate with these studies as we do use AI for projects and assignments. And with this we got another aspect about plagiarisms which are mandated by professors and universities to get an understanding of the authenticity of work submitted by students but since we are habitual to follow the responses of AI with 100% trust we do fail plagiarism thresholds and face difficulties to get the work to acceptance level which flags about the damage to cognitive engagement and critical thinking. These are the signs that shows we are at a phase where we can’t stop students from using AI even if they are being restricted with rules by professor or the institution. What is your critical analysis of this situation? To have the real understanding of how AI is being used we have initiated student surveys for analysis. We have prepared a very intensive survey questionnaire which completely relates to student life. We have initiated this survey drive in various universities for better perspective. The survey has been initiated in Pace University which is very well known for its profession-focused approach, New York University, which is private large-scale university known for its research oriented methods for teaching and learning, Columbia University to get insights about one of the best Ivy League institutions, colleges from City University of New York which is a premier public university known for high quality education like Baruch college, Hunter college, City college of NY, Queens college. As per the responses we received, students have share concerns about accuracy of the responses due to hallucinations and also about the security as that's always a major factor whenever anything gets to market and is in use. What perspectives were discussed and how were they debated or arbitrated within your team? A collaboration between 2 universities and different backgrounds did open doors to many new perspectives during brainstorming sessions. Our team discussed different perspectives on the use of Al by students. We agreed that Al has become necessary not only for education, but also for daily life. From using AI for eduction to asking about horoscope or plan a trip. However, we also discussed the risks of Al use, including overdependence, hallucination, and privacy concerns. Some team members were concerned that students might rely too much on Al and loose cognitive element. Others believed that preventing students from using Al is not realistic, because Al is already part of modern life and in some way acts as addiction. After discussion, our team agreed that the best approach is not to stop students from using Al, but to encourage them to use it in the right way. This means using Al as a support tool while making sure students still gain knowledge from assignments and projects, avoid errors caused by hallucinated information, and keep their personal data safe.While we had a focus on students learning and use of AI, we also came across thoughts about how this AI partnership can be implemented from a point of government imposing policy to AI companies accepting and implementing with no loss of business but having the revenue increased. We came across the idea to have LLMs licensed for educational purposes for students in partnership with universities and educational institutions where these institutions can have a separate environment for students partnered with AI companies to support ethical use. This can help institutes to understand the creativity and innovative thinking of students in give right direction for future. What contribution are you proposing, and under what conditions could it be implemented? We would like to propose the ethical AI partnership where all stakeholders (students, AI companies, educational institutions, government) participate mutually. Through our research and survey and got attention of 3 major concerns: cognitive engagement reduction, AI hallucination and privacy issues. For growth and future of students in this era a free hand for use of AI is very important to see what they can come up with out of the box ideas with the power of AI. For this we propose students to use AI and share the prompts used by them along with the work (assignment/project) they submit. We would like to have LLMs not to answer the questions asked straight away but collaborate with students for their learning without any loss of cognitive thinking and skills by helping them to learn. Another major concern is AI hallucination, which remains a serious challenge across the industry and a solution to resolve this can be time taking and complex. One practical step is to require AI system to provide sources or references alongside their response whenever possible which can help to reduce risk of misleading and false responses and this partnership with AI can help students to learn from references and validate the responses. Finally, privacy and data security is always a concern in the digital space and with AI the guards have to be always up to be safe and methods like zero trust policy are really helping but AI companies must take greater responsibility for protecting student data. Educational institutions and government agencies should only allow the use of AI platforms that have passed privacy and data protection audits, such as ISO/IEC 27701:2025 for Privacy Information Management Systems (PIMS). References 1. Hassen, M. Z. (2025). The impact of AI on students’ reading, critical thinking, and problem-solving skills. American Journal of Education and Information Technology, 9(2), 82-90. 2. Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X. H., Beresnitzky, A. V., ... & Maes, P. (2025). Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task. arXiv preprint arXiv:2506.08872, 4.