The study (found by Kevin), “AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking,” examines how increasing reliance on AI tools affects users’ critical thinking abilities, with a particular focus on the concept of cognitive offloading. Using a mixed-methods approach with surveys and interviews from over 650 participants across different age groups, the researchers found a significant negative relationship between frequent AI tool usage and critical thinking skills, especially among younger users. The researchers argue that when individuals consistently delegate cognitive work, such as information retrieval or decision-making, to AI systems, they engage less deeply with problems themselves. This study is important to our problem space because it establishes that over-reliance on AI can weaken critical thinking, setting the foundation for further investigation into why some users rely too heavily on these tools and how that reliance might be mitigated.
Expanding on this broader societal framing, the next study (found by Kevin), “AI Overdependence in Education: Examining Its Harmful Effects on Student Learning and Critical Thinking,” narrows the focus specifically to educational contexts. This paper explores how frequent AI use in academic settings can lead to passive learning behaviors, reduced problem-solving skills, and diminished critical engagement. Using measures such as the Critical Thinking Assessment Scale, the researchers show that students who rely heavily on AI-generated content are less likely to analyze, evaluate, or verify information independently. Rather than rejecting AI outright, the authors finally emphasize the need for balance between AI assistance and traditional learning approaches. This study directly connects to other research we’ve cited by reinforcing the idea that the manner in which students rely on AI, not just the frequency of use, plays a key role in shaping learning outcomes.
Connecting this discussion to students’ lived experiences, the third study (found by Kevin), “Student Perspectives on the Benefits and Risks of AI in Education,” captures how undergraduate students themselves perceive AI tools. Through a survey of 262 students at a large U.S. university, the researchers identified perceived benefits such as efficiency, personalized support, and access to information, alongside risks including loss of critical thinking, academic integrity concerns, and over-reliance. Many students expressed anxiety about trusting AI outputs too readily without understanding their limitations. This study is particularly useful as a bridge to other research, as it highlights students’ awareness of over-reliance as a risk while also revealing uncertainty around what healthy AI use actually looks like.
Extending on factors that shape use in practice, the study (found by Delali), “Students’ Reliance on AI in Higher Education: Identifying Contributing Factors,” was conducted with the goal of understanding what factors contribute most to overreliance, healthy reliance, and underreliance on AI by students in high eduction. Research was conducted with 50 undergraduate students who were enrolled in computer science related courses and had prior programming experience. They were given 14 Python programming tasks and asked to solve them with an AI assistant which provided both accurate and intentionally inaccurate suggestions. They found that there was a correlation between appropriate reliance on AI and students’ own belief in their capacity to solve the task as well as the extent of their programming skills and knowledge. This study is relevant to our problem space as it grapples with what is a healthy level of reliance on AI for university students. We seek to craft a solution that teaches students what appropriate reliance looks like.
Finally, building on this focus on healthy versus unhealthy reliance, in another study (found by Delali), “The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: a systematic review,” this research is taken further in order to investigate how overreliance directly affects students’ cognition, particularly their decision making, critical thinking, and reasoning. In this article, over-reliance is defined as accepting AI-generated recommendations without question. The systematic literature review encompasses 14 other research articles, placing a particular focus on the ethical conversation related to AI chat-based systems. The review found that over-reliance on AI does indeed affect cognitive abilities, because students’ favor fast solutions over slower, practical ones. They recommend that encouraging students to reflect on ethics and risks associated with trusting AI recommendations can cause more critical thinking and usage with these tools. This study is particularly interesting to us because of its recommendations around how to reduce overreliance by encouraging students to take the ethics framing approach.
