Educational AI and Development of Critical Thinking and Cognition

This work examines how digital technologies and educational AI influence critical thinking and cognition, proposing a framework for integrating generative AI into pedagogical practice. Within educational contexts, AI is restructuring pedagogy and administration, yet resource abundance often induces cognitive overload. Contemporary research reveals the reverse Flynn effect—declining IQ and cognitive abilities, particularly among Generation Z, observable since approximately 2010. Yale University reports a 90% increase in cognitive difficulties among adults aged 18–34 between 2013 and 2023, with declines in verbal, numerical, and fluid reasoning. Excessive screen exposure is a primary factor, with students using computers five hours daily showing lower performance in attention, memory, literacy, and executive function. Cognitive offloading—relying on digital tools to reduce mental effort—impairs long-term memory formation, with a strong negative correlation between offloading frequency and recall accuracy. The "Google effect" exemplifies this, as individuals remember information location rather than content, reducing hippocampal activity. While offloading can be beneficial when mental energy is redirected to deeper understanding, passive use creates cumulative cognitive deficits. Print reading generates stronger neural connectivity in language and emotion-processing regions than screen reading, which reduces hippocampal activation. Screen readers multitask and scan, while print readers sustain attention; a 2024 meta-analysis confirms print superiority for comprehension across 49 studies. Chronic digital access triggers neuroplastic restructuring that favors rapid task-switching over deep processing. These challenges demand solutions beyond content organization. The OECD's Digital Education Outlook 2026 reports that students using general-purpose AI improved short-term performance by 48% but performed 17% worse without access, confirming that learning gains are unsustainable when technology replaces cognitive effort. Generative AI offers opportunities for personalized scaffolding but risks cognitive offloading and skill atrophy. While AI expedites early learning, it may bypass the cognitive struggle essential for internalizing core procedures, compromising long-term retention and transfer. AI excels at replication but lacks human capacity for ethical judgment and meaningful integration. Effective pedagogy requires metacognitive instruction, process documentation, and assessments rewarding deep engagement over superficial output. Adaptive platforms enforcing active engagement and desirable difficulties demonstrate superior retention compared to static study aids. The future depends on designing AI as a tutor promoting reasoning rather than a solver providing answers, ensuring the human remains cognitive driver and AI serves as cognitive gym strengthening mental capacity. This alignment can transform AI from a source of cognitive decay into an engine for cognitive growth, provided pedagogical innovation prioritizes deep engagement, metacognitive awareness, and gradual autonomy development. The evidence presented demonstrates that AI has potential as a powerful engine for cognitive growth, but only when used strategically to support, rather than replace, the cognitive work of learning. The distinction between strategic augmentation and passive delegation, the difference between extraction and synthesis, and the impact of interface design on schema consolidation all point to a single conclusion: the quality of learning in the age of AI depends fundamentally upon the intentionality of instructional design and the preservation of human cognitive agency.