Is AI Enhancing Education or Hurting It?

Artificial intelligence, particularly large language models such as ChatGPT, has rapidly entered educational settings. Students now use AI tools to summarize readings, generate explanations, brainstorm ideas, and practice problem solving. Educators and healthcare professionals alike are asking an important question: do these tools support human cognition, or do they weaken the very skills education is meant to develop? Cognitive psychology provides a useful framework for evaluating both the benefits and risks of AI-assisted learning.

Topic Summary and Theory Background

Research on large language models and cognition has expanded quickly because these tools directly interact with core cognitive processes including memory, attention, language comprehension, and problem solving. Unlike earlier technologies that primarily stored information, LLMs actively generate explanations, feedback, and structured responses, making them more cognitively interactive. As a result, researchers are examining how AI influences the way learners encode information, retrieve knowledge, and allocate cognitive effort during learning tasks.

Evidence suggests that AI tools can reduce cognitive load by offloading lower-level tasks such as summarization or factual retrieval. This may allow learners to focus more attention on higher-order thinking, including synthesis and application. However, concerns remain that frequent reliance on AI could reduce deep processing, leading to weaker long-term retention and poorer transfer of knowledge. Similar concerns emerged in earlier research on search engines and navigation technologies, which were shown to alter memory strategies and problem solving over time.

One foundational cognitive framework relevant to AI-assisted learning is levels of processing theory. This theory proposes that information processed at a deeper semantic level is remembered better than information processed shallowly. When students actively generate explanations or struggle through problems, they engage in deeper processing. If AI provides immediate answers, learners may remain at a shallow level of engagement. However, when AI is used as a scaffold, such as prompting reflection or offering feedback, it may support deeper processing rather than replace it.

Comparing Primary Research on Technology and Cognition

Research on AI and cognition builds on earlier work examining how external technologies influence thinking. Two influential studies in this area are Sparrow et al. (2011) and Barr et al. (2015), with more recent educational AI research summarized by Zhai et al. (2023). Together, these sources illustrate how technology can both support and alter cognition depending on how it is used.

Sparrow et al. (2011) conducted laboratory-based experiments with adult participants to examine how access to search engines affects memory. Participants completed trivia tasks while believing information would either be saved or unavailable later. In contrast, Barr et al. (2015) studied a broader adult population and focused on real-world smartphone use, measuring how device reliance related to analytic thinking and cognitive reflection. While Sparrow et al. emphasized experimental control, Barr et al. captured more naturalistic patterns of technology reliance. This contrast highlights a key challenge for AI research in education: laboratory findings may not fully reflect how students actually use AI tools in everyday learning environments.

The studies also differed in research design and measurement. Sparrow et al. (2011) used experimental manipulation and objective memory measures, allowing for stronger causal conclusions about how external information storage affects encoding. Barr et al. (2015) relied on correlational designs using self-report measures and performance on cognitive reflection tasks. While this limits causal inference, it provides insight into habitual technology use and thinking styles. Zhai et al. (2023) reviewed educational AI studies using both approaches, noting that experimental designs often show short-term cognitive benefits from AI scaffolding, while observational studies raise concerns about overreliance and reduced deep processing.

Each study also identified important limitations. Sparrow et al. (2011) focused on short-term memory effects, leaving questions about long-term learning unanswered. Barr et al. (2015) could not determine whether technology use caused reduced analytic thinking or whether individuals with lower analytic engagement were more likely to rely on devices. Zhai et al. (2023) further noted that many AI-in-education studies rely on convenience samples and rapidly evolving tools, limiting generalizability. Together, these findings suggest that AI can function as cognitive support, but its long-term effects remain unclear without longitudinal research.

Conclusions

Based on current evidence, artificial intelligence is neither inherently enhancing nor inherently harming education. Its cognitive impact depends largely on how it is used. When AI replaces active engagement, it risks encouraging shallow processing and overreliance, potentially weakening memory and problem solving. When used as a supportive tool that encourages reflection, elaboration, and metacognition, AI can enhance learning and reduce unnecessary cognitive load.

From a cognitive psychology perspective, effective educational use of AI aligns with established principles such as levels of processing and cognitive load theory. Educators and healthcare professionals should focus on teaching students how to use AI strategically as a supplement rather than a substitute for thinking. As AI continues to shape learning environments, grounding its use in cognitive science will be essential to ensure these tools strengthen, rather than erode, human cognition.

References

Barr, N., Pennycook, G., Stolz, J. A., & Fugelsang, J. A. (2015). The brain in your pocket: Evidence that smartphones are used to supplant thinking. Computers in Human Behavior, 48, 473–480.

Sparrow, B., Liu, J., & Wegner, D. M. (2011). Google effects on memory: Cognitive consequences of having information at our fingertips. Science, 333(6043), 776–778.

Wegner, D. M. (1987). Transactive memory: A contemporary analysis of the group mind. In B. Mullen & G. R. Goethals (Eds.), Theories of group behavior (pp. 185–208). Springer.

Zhai, X., Wang, M., & Ghani, U. (2023). The impact of artificial intelligence on learning: A review of educational AI research. Computers & Education: Artificial Intelligence, 4, 100126.