Harold Robert Meyer | The ADD Resource Center
haroldmeyer@addrc.org http://www.addrc.org/
Reviewed 04/11/2026 – Published 05/03/2026
Listen to understand, not just to respond
(With help from Claude, Gemini, Gork, and Perplexity)
Executive Summary
As AI tools become embedded in daily life, emerging research reveals a troubling tradeoff: the more we outsource our thinking to machines, the less we exercise the cognitive muscles that keep us sharp. This phenomenon—known as cognitive offloading—has particular implications for individuals with ADHD, who already navigate executive function challenges and may be both the greatest beneficiaries and the most vulnerable users of AI-powered tools. The research is converging on a clear message: the issue is not whether to use AI, but whether we can cultivate the judgment to use it without surrendering the thinking that makes us who we are.
Why This Matters
A March 2026 RAND report found that 62% of middle school, high school, and college students now use AI for homework—up from 48% just seven months earlier. Even more striking, 67% of those students believe that the more they use AI for schoolwork, the more it will harm their critical thinking skills. They’re using it anyway.
For the ADHD community, this tension is magnified. AI tools offer genuinely transformative support—scaffolding executive function, reducing cognitive load, and helping with task initiation. But the same cognitive offloading that provides relief may, over time, weaken the very skills individuals with ADHD most need to develop and maintain.
Key Findings
The Science of Cognitive Offloading
Cognitive offloading—delegating mental tasks to external tools—is not new. We’ve done it with writing, calculators, and calendars for centuries. But AI represents a qualitative leap. Unlike a calculator that handles arithmetic, generative AI can now perform analysis, reasoning, and creative composition—cognitive functions that were once exclusively human territory.
A landmark 2025 study by Michael Gerlich at SBS Swiss Business School surveyed 666 participants and found a statistically significant negative correlation (r = −0.68) between frequent AI tool usage and critical thinking scores. The effect was mediated by cognitive offloading: the more participants relied on AI, the less they engaged their own analytical faculties. Cognitive offloading was strongly correlated with AI usage (r = +0.72) and inversely related to critical thinking (r = −0.75). Younger users, ages 17–25, showed the highest AI dependence and the lowest critical thinking scores, while higher educational attainment appeared to buffer against the worst effects.
A separate 2025 study published at the CHI Conference, led by Microsoft Research, surveyed 319 knowledge workers who shared 936 real-world examples of using generative AI in workplace tasks. The study found that generative AI consistently reduced the perceived effort required for critical thinking. Workers who had high confidence in AI’s capabilities were especially likely to scale back their own cognitive engagement. The study documented a notable shift from “material production to critical integration”—workers were spending less time doing the thinking and more time reviewing AI output, but the quality of that review often depended on their pre-existing expertise. Critically, workers with higher self-confidence in their own domain knowledge maintained stronger critical thinking practices even when using AI regularly.
A 2026 study by Yurt and Kuşci in Current Psychology identified the specific psychological mechanisms at work. Surveying 422 university students, they found that AI tool usage significantly predicted both “epistemic laziness”—a reduced inclination to seek out and verify information independently—and “metacognitive weakness,” a diminished capacity to monitor and regulate one’s own thinking. Both of these factors mediated the relationship between AI use and diminished critical thinking, suggesting that AI doesn’t just make thinking easier—it can erode our motivation and ability to think about thinking.
A study by Gonsalves (2024) in the Journal of Marketing Education proposed a revised Bloom’s Taxonomy framework that accounts for AI-specific competencies. Tracking MSc Marketing students’ interactions with AI tools over four weeks, the study found that AI can both enhance and challenge critical thinking—but only when students engage metacognitively, thinking about their AI interactions rather than passively accepting outputs. Students who moved fluidly between cognitive stages—analyzing and evaluating simultaneously—showed the strongest learning outcomes.
The Memory Paradox
Cognitive psychologist Barbara Oakley and a team of neuroscience researchers identified what they call “The Memory Paradox”: as AI tools grow more capable, our brains increasingly opt out of the effortful processing that builds durable knowledge. When we repeatedly retrieve information from our own memory, we form robust neural pathways—schemas that support intuition, pattern recognition, and error-checking. When we default to “just ask the AI,” those pathways may never fully develop.
This echoes what researchers have observed with other technologies. Neuroscientist Louisa Dahmani’s research, published in Scientific Reports, assessed 50 regular drivers and found that habitual GPS users developed weaker spatial memory and reduced hippocampal engagement compared to people who navigated using their own cognitive maps. In a longitudinal follow-up, greater GPS use over a three-year period was associated with a steeper decline in hippocampal-dependent spatial memory. Crucially, those who used GPS more did not do so because they felt they had a poor sense of direction—suggesting that the technology drove the decline, not the other way around.
The parallel to AI is direct. If generative AI handles your writing, planning, analysis, and problem-solving, the brain regions and cognitive processes responsible for those functions get less exercise. Research on “metacognitive laziness” by Fan et al., published in the British Journal of Educational Technology, found that while students using ChatGPT produced better essays in the short term, they showed no advantage in actual knowledge gain or transfer—and exhibited reduced engagement in the self-regulated learning processes (planning, monitoring, evaluating) that build lasting competence.
The trajectory of rising IQ scores observed from the 1930s through the 1980s—known as the Flynn effect—has plateaued and even reversed in several developed countries, a trend some researchers attribute partly to increased cognitive offloading. Research by Meyerhoff et al. (2021), published in Psychonomic Bulletin & Review, demonstrated the core tradeoff experimentally: cognitive offloading boosts immediate task performance but diminishes memory formation, confirming that the convenience of external tools comes at a measurable cognitive cost.
The Risk: Cognitive Atrophy and “Human Imbecilment”
Some researchers have moved beyond polite warnings to a blunter assessment. Neuropsychologist Umberto León Domínguez of the University of Monterrey has argued that the cumulative effect of AI dependence amounts to what might fairly be called human imbecilment—a progressive dulling of the very cognitive abilities that define independent thought. Unlike earlier tools that handled one narrow task, generative AI can now plan, reason, and compose from start to finish—leaving no cognitive muscle for the user to flex. The risk, Domínguez warns, is not merely laziness but a structural weakening of higher-order executive functions, including precisely the planning, decision-making, and problem-solving skills that individuals with ADHD most need to develop and maintain. He describes AI as a “cognitive prosthesis” that, when overused, doesn’t just assist the brain—it begins to replace it.
When we outsource the heavy lifting of logic and synthesis, we risk losing the mental muscle memory required for independent thought. Three specific mechanisms drive this risk:
The “Easy Button” Syndrome. When AI summarizes a report instead of you reading it, you miss the nuances, the biases, and the gaps. You aren’t learning the material—you’re learning how to prompt. The cognitive engagement that builds understanding is entirely bypassed.
Acceptance Bias. We have a natural tendency to trust authoritative-sounding text. Since AI is engineered to produce confident, fluent prose—even when generating fabricated content—users often stop fact-checking. This creates a form of “truth decay” where the appearance of expertise substitutes for its reality.
Algorithmic Echo Chambers. AI models are trained on existing human data. Relying on them exclusively can trap us in a loop of statistically average ideas, stifling the radical, outside-the-box thinking that defines human breakthrough. The very creativity and divergent thinking that individuals with ADHD often possess could be flattened by over-reliance on AI-generated frameworks.
The Reward: Cognitive Augmentation
On the flip side, AI can act as a powerful Socratic partner that pushes thinking further than we could go alone—when used intentionally.
The Ultimate Sounding Board. You can use AI to play devil’s advocate. Asking “What are the flaws in my argument?” or “Give me three counterpoints to this thesis” actively engages your critical faculties rather than replacing them.
Information Synthesis. AI can process vast amounts of data in seconds. Used correctly, this frees up mental bandwidth for high-level strategy and creative problem-solving rather than rote data sorting. The key is using AI to handle what neuroscientists call “extraneous cognitive load”—the background noise of a task—while preserving your engagement with “germane load,” the actual learning and deep thinking.
Drafting vs. Thinking. Using AI to overcome blank-page syndrome allows you to spend more time editing and refining, which are themselves high-level critical thinking skills. The cognitive engagement of evaluating, restructuring, and making AI output your own is where the real thinking happens.
What This Means for ADHD
For individuals with ADHD, the relationship with AI tools is genuinely paradoxical. Executive function coaching expert Yulia Rafailova describes AI as a “cognitive scaffold” that can reduce the crushing weight of working memory demands, support task initiation, and help with planning and prioritization. For ADHD brains that struggle with activation energy—the mental push needed to begin a task—AI can bypass the freeze response and provide a starting structure.
But Rafailova also warns that over-reliance on AI, particularly from a young age, could weaken capacity for creativity, problem-solving, and resilience. The distinction she draws is critical: using AI as a scaffold (a temporary support while building skills) versus a crutch (a permanent replacement for those skills).
Russell Barkley, a leading ADHD researcher, has long argued that ADHD is fundamentally a disorder of self-regulation rooted in executive function deficits—specifically working memory, cognitive flexibility, and inhibitory control. These are precisely the cognitive functions most at risk from chronic offloading. If an individual with ADHD never practices planning because AI does it for them, or never struggles through organizing thoughts because a language model does the composing, those already-challenged neural pathways receive even less reinforcement.
The neuroscience framing clarifies the ideal balance. For the ADHD brain, AI can handle extraneous load—formatting, organizing, information retrieval—freeing up cognitive resources for germane load, the actual learning and deep thinking. The risk is when AI starts handling the germane load too, and the brain’s engagement with challenging material drops to near zero.
The Genie Is Out of the Bottle
There is no realistic scenario in which AI tools become less prevalent, less capable, or less accessible. Historical parallels abound: the printing press sparked fears of declining memory; calculators raised concerns about math skills. In each case, society adapted by elevating higher-order skills—analysis, synthesis, and ethical judgment.
The question is not whether people will use AI—they already do and increasingly will. The question is whether we can develop the metacognitive awareness to direct and scrutinize it, rather than passively accepting its outputs.
The research identifies several protective factors. Higher education and domain expertise buffer against the worst effects of cognitive offloading. Workers confident in their own knowledge apply more critical scrutiny to AI outputs. And intentional practice—deliberately choosing when to engage your own thinking versus when to delegate—may preserve cognitive function over time.
The most valuable skill in 2026 isn’t knowing the answer—it’s knowing how to ask the right questions and having the discernment to judge the response. — Harold Meyer
Balancing the Scales: Passive Reliance vs. Active Integration
The difference between AI as a crutch and AI as a bicycle for the mind comes down to how you use it:
Problem Solving. Passive: Asking AI for “the answer.” Active: Asking AI for “five frameworks to approach this problem,” then choosing and adapting the best one yourself.
Writing. Passive: Letting AI generate the final draft. Active: Using AI to outline or overcome a blank page, then writing the substance yourself and using AI to refine it.
Fact-Checking. Passive: Assuming the output is correct. Active: Treating AI output as a starting hypothesis for manual verification against primary sources.
Learning. Passive: Asking for a summary. Active: Asking AI to quiz you on the material, challenge your understanding, or identify gaps in your reasoning.
Practical Strategies for the ADHD Community
Use AI as a launchpad, not a landing pad. Let AI generate a first draft or a task breakdown, then do the thinking work of evaluating, revising, and making it your own. The cognitive engagement of refining AI output is where critical thinking happens.
Practice selective offloading. Not every task warrants the same level of cognitive investment. Use AI freely for low-stakes administrative tasks, but deliberately engage your own thinking for decisions, creative work, and anything that involves your values or judgment.
Build in retrieval practice. Periodically challenge yourself to recall information, make plans, or solve problems without AI assistance. The act of retrieval—pulling information from your own memory rather than an external source—strengthens neural pathways and supports long-term retention.
Treat AI output with healthy skepticism. AI produces confident-sounding text whether or not it’s accurate. Make it a habit to fact-check, question assumptions, and consider alternative perspectives before accepting AI-generated content. This verification process is itself a critical thinking exercise.
Use AI to ask better questions. Rather than asking AI for answers, try asking it to help you think through a problem. Prompt it to challenge your reasoning, offer counterarguments, or identify what you might be missing. This shifts the dynamic from passive consumption to active engagement.
Monitor your own patterns. Notice when you reach for AI out of genuine need versus when you reach for it out of avoidance or convenience. For individuals with ADHD, the impulse to offload discomfort is strong—but the discomfort of cognitive effort is often where growth happens.
Create AI-free zones. Schedule regular sessions for reading, writing, or problem-solving without any AI tools. Build the mental muscle the way you’d build physical muscle—through deliberate, effortful practice.
“AI doesn’t replace your thinking—it reveals how much of your thinking you’re willing to do yourself.” — Harold Meyer
The Bottom Line
AI tools offer extraordinary benefits for individuals with ADHD—reducing cognitive load, supporting executive function, and making complex tasks more manageable. But these benefits come with a cognitive cost that deserves honest acknowledgment. The research is increasingly clear: frequent, uncritical reliance on AI is associated with diminished critical thinking, weakened memory formation, reduced independent analytical capacity, and the emergence of epistemic laziness and metacognitive weakness. Left unchecked, the result is what researchers are beginning to recognize as human imbecilment—not a sudden collapse, but a slow, comfortable slide toward cognitive dependency.
The genie is indeed out of the bottle. We cannot and should not try to stuff it back in. But we can—and must—develop the metacognitive awareness to use AI intentionally, maintaining the cognitive engagement that keeps our minds active and our critical faculties sharp. For the ADHD community, the stakes are especially high, and the balance especially important.
The goal is not to use AI less, but to use it better.
Related ADDRC Resources
ADHD and Executive Functioning: https://www.addrc.org/what-is-the-difference-between-adhd-and-executive-function-disorder/
ADHD and Decision-Making: Why It’s So Hard and What Helps: https://www.addrc.org/practical-strategies-for-managing-decision-making-with-adhd/
ADHD and Overwhelm: Practical Strategies for Managing Cognitive Load: https://www.addrc.org/adhd-and-overwhelm-why-it-hits-harder-and-what-to-do/
ADHD and Technology: Tools That Help (and When They Don’t): https://www.addrc.org/understanding-adhd-in-the-digital-age-practical-strategies-for-success/
ADHD in the Workplace: Advocating for Yourself: https://www.addrc.org/how-to-advocate-for-yourself-at-work-with-adhd/
References
Dahmani, L., & Bohbot, V. D. (2020). Habitual use of GPS negatively impacts spatial memory during self-guided navigation. Scientific Reports, 10, 6310. https://doi.org/10.1038/s41598-020-62877-0
Domínguez, U. L. (2024). ChatGPT as a cognitive prosthesis: Implications for higher-order executive functions. Neuropsychology, 38(2), 104–112. https://doi.org/10.1037/neu0000938
Fan, Y., Tang, L., Le, H., Shen, K., Tan, S., Zhao, Y., Shen, Y., Li, X., & Gašević, D. (2025). Beware of metacognitive laziness: Effects of generative artificial intelligence on learning motivation, processes, and performance. British Journal of Educational Technology, 56, 489–530. https://doi.org/10.1111/bjet.13544
Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6. https://doi.org/10.3390/soc15010006
Gonsalves, C. (2024). Generative AI’s impact on critical thinking: Revisiting Bloom’s Taxonomy. Journal of Marketing Education. https://doi.org/10.1177/02734753241305980
Lee, H.-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. In CHI Conference on Human Factors in Computing Systems (CHI ’25). ACM. https://doi.org/10.1145/3706598.3713778
Meyerhoff, H. S., Grinschgl, S., Papenmeier, F., & Gilbert, S. J. (2021). Consequences of cognitive offloading: Boosting performance but diminishing memory. Psychonomic Bulletin & Review, 28, 1474–1486. https://doi.org/10.3758/s13423-021-01916-x
Oakley, B., et al. (2025). The Memory Paradox. [Referenced in Policy Options, September 2025.]
Schwartz, H. L., & Diliberti, M. K. (2026). More students use AI for homework, and more believe it harms critical thinking: Selected findings from the American Youth Panel. RAND Corporation, RR-A4742-1. https://www.rand.org/pubs/research_reports/RRA4742-1.html
Yurt, E., & Kuşci, I. (2026). Factors influencing critical thinking during AI use among university students: The mediating effects of epistemic laziness and metacognitive weakness. Current Psychology, 45, 67. https://doi.org/10.1007/s12144-025-08800-0
—
About The Author
Harold Meyer is the founder of The A.D.D. Resource Center, established in 1993. For over 30 years, he has been a leading advocate, coach, and educator in the ADHD space, translating the real experiences of individuals with ADHD into practical guidance for families, professionals, and institutions. He co-founded CHADD of New York, served as CHADD’s national treasurer, and was president of the Institute for the Advancement of ADHD Coaching. An author and international speaker, he has presented at the American Psychiatric Association and CHADD national conferences. haroldmeyer@addrc.org
Contact
info@addrc.org (mailto:info@addrc.org) • +1 (646) 205-8080
127 West 83rd St., Unit 133, Planetarium Station, New York, NY 10024-0840 USA
X | LinkedIn | Substack | ADHD Research and Innovation
Join Our Community
Subscribe to the ADD Resource Center newsletter for the latest resources and insights → Click here.
Disclaimers
Content is for educational purposes only and is not a substitute for professional advice. We strive for accuracy, though errors can occur. Some material may be AI-generated; please verify independently. Rejection Sensitive Dysphoria (RSD) is recognized by many providers but is not in the DSM.
In the USA and Canada, call or text 988 anytime for free mental health and suicide prevention support.
Privacy & Legal
Under GDPR and CCPA, you may request access to, correction of, or deletion of your personal data at info@addrc.org.© 2026 Harold R. Meyer / ADD Resource Center. All rights reserved. Content may be shared only in complete, unaltered form with attribution. Reproduction or commercial use requires written permission (addrc@mail.com).
The ADD Resource Center: Your Trusted Source for ADHD for ADHD information and research. Practical strategies. Expert guidance—for people with ADHD and everyone in their world.