{"id":4870,"date":"2025-07-16T15:43:38","date_gmt":"2025-07-16T13:43:38","guid":{"rendered":"https:\/\/matbury.com\/wordpress\/?p=4870"},"modified":"2025-08-28T14:58:45","modified_gmt":"2025-08-28T12:58:45","slug":"think-before-you-rely-on-chatgpt-advice-for-students-using-an-llm-to-study","status":"publish","type":"post","link":"https:\/\/matbury.com\/wordpress\/index.php\/2025\/07\/16\/think-before-you-rely-on-chatgpt-advice-for-students-using-an-llm-to-study\/","title":{"rendered":"Think Before You Rely on ChatGPT: Advice for students using an LLM to study"},"content":{"rendered":"<p><a href=\"https:\/\/matbury-com.translate.goog\/wordpress\/index.php\/2025\/07\/16\/think-before-you-rely-on-chatgpt-advice-for-students-using-an-llm-to-study\/?_x_tr_sl=en&amp;_x_tr_tl=es&amp;_x_tr_hl=en-US&amp;_x_tr_pto=wapp\" target=\"_blank\" rel=\"noopener\">Haga clic aqu\u00ed para la traducci\u00f3n al espa\u00f1ol (Google Translate)<\/a><\/p>\n<p class=\"western\">The following is my response to the current discourse surrounding a recent pre-publication paper, \u201c<i>Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task<\/i>\u201d (Kosmyna et al., 2025) from a study that examined &amp; compared the effects of students using a large language model (LLM) to help them write essays vs unassisted essay writing. Although the paper has yet to go through peer-review, I think its conclusions concur with many predictions &amp; conjectures that science of learning researchers have made about the potential effects of LLM use by students, &amp; that it\u2019s justifiable to publish this article offering helpful, evidence-informed advice.<\/p>\n<h3 class=\"western\">Introduction<\/h3>\n<p class=\"western\"><b>These days, it\u2019s easy to turn to tools like ChatGPT or Google Gemini when you\u2019re stuck on an assignment, short on time, or just looking for inspiration. To be honest, it can be incredibly helpful. Whether it\u2019s summarising a topic, correcting grammar, or giving you a structure to start with, using an LLM can feel like a smart shortcut.\u00a0<\/b><b>But here\u2019s something you might not have considered: using an LLM too often or in the wrong way could actually inhibit your learning. Recent research shows that too heavy a reliance can weaken memory, creativity, and critical thinking; the very skills you need to succeed, not just in school, but in life.\u00a0<\/b><b>So, before you let an LLM write your next essay or explain every concept for you, read on. What follows is advice to help you use them wisely so you can accelerate your learning rather than hold it back.<\/b><\/p>\n<h3 class=\"western\">Overuse of LLMs can hurt your brain\u2019s growth<\/h3>\n<p class=\"western\"><b>Cognitive shortcuts result in cognitive costs<\/b><\/p>\n<p class=\"western\">When you let an LLM do the heavy thinking for you, obviously, your brain stops working as hard. This &#8220;cognitive offloading&#8221; may feel helpful in the moment, but over time, it can lead to weaker critical thinking, reduced creativity, &amp; greater mental passivity. You might get used to the ease and find yourself struggling when you have to think independently.<\/p>\n<p class=\"western\"><b>Useful analogy<\/b>: Just as we work out at the gym to strengthen muscles, we do academic writing to strengthen our brains. What effect do you think it would have if you reduced the weights &amp; resistance at the gym by about 50%? That\u2019s a similar effect to LLM-assisted essay writing!<\/p>\n<p class=\"western\"><b>Brainpower declines with dependence<\/b><\/p>\n<p class=\"western\">What\u2019s more, according to the paper, students who used LLMs repeatedly showed reduced brain activity in regions linked to concentration, memory, &amp; decision-making. When they later had to write without an LLM, their performance &amp; brain engagement dropped. Why? Their minds had grown dependent on the LLM &amp; that made it harder to think without help.<\/p>\n<h3 class=\"western\">Shallow input results in shallow learning<\/h3>\n<p class=\"western\"><b>You learn less when you rely on instant answers<\/b><\/p>\n<p class=\"western\">It\u2019s tempting to ask ChatGPT for a quick explanation or essay paragraph but if you don\u2019t think through the material yourself, asking questions, making mistakes, &amp; revising ideas, you won\u2019t remember it well. Students who used an LLM were worse at remembering their own writing &amp; often couldn\u2019t quote their own essays correctly, indicating that they hadn\u2019t really learnt much about the topics.<\/p>\n<p class=\"western\"><b>Less thinking, more tweaking<\/b><\/p>\n<p class=\"western\">Instead of developing their own ideas, students using LLMs spent more time editing or approving what the LLM wrote. That means fewer opportunities to practise fundamental academic skills like reasoning, brainstorming, &amp; synthesising knowledge.<\/p>\n<h3 class=\"western\">The brain works differently when an LLM is involved<\/h3>\n<p class=\"western\"><b>Reduced creativity &amp; self-organisation<\/b><\/p>\n<p class=\"western\">According to the study, LLM-assisted writing led to lower activity in brain areas tied to creative thought &amp; self-organisation. Rather than generating original ideas, students often just filtered what the tool gave them, resulting in more generic, less dynamic writing.<\/p>\n<p class=\"western\"><b>LLMs alter how you process information<\/b><\/p>\n<p class=\"western\">When students wrote without LLM assistance, their brains showed &#8220;bottom-up&#8221; processing; where ideas are generated from students\u2019 own memories, which is how our brains learn best. With an LLM, that flipped: brains were in &#8220;top-down&#8221; mode, mostly reacting to external input instead of generating fresh ideas internally.<\/p>\n<h3 class=\"western\">You may lose your voice, literally<\/h3>\n<p class=\"western\"><b>Weaker sense of authorship<\/b><\/p>\n<p class=\"western\">Students relying on an LLM often reported a vague or incomplete sense of ownership over their essays. The words felt less like theirs simply because they weren\u2019t. This loss of ownership can reduce your motivation, self-confidence, &amp; pride in your work.<\/p>\n<p class=\"western\"><b>Homogeneity &amp; bias<\/b><\/p>\n<p class=\"western\">LLM-written essays tend to sound the same; structured, safe, &amp; uninspired. Teachers can usually spot them. They also reflect the biases of the data that the LLMs were trained on. If you don\u2019t critically engage with the content for yourself, your writing risks becoming dull, predictable, &amp; even misinformed.<\/p>\n<h3 class=\"western\">So, what should you do instead?<\/h3>\n<p class=\"western\">It\u2019s not necessarily about avoiding LLMs altogether, although that\u2019s the simplest, safest option, it\u2019s about using it purposefully, sparingly, &amp; wisely.<\/p>\n<p class=\"western\"><b>Start without an LLM first<\/b><\/p>\n<p class=\"western\">In the early stages of learning or writing, build your ideas without LLM assistance. Let your brain do the work. It\u2019s more difficult but this lays down strong mental connections that will serve you in exams, discussions, &amp; real-world problem-solving. This is what Robert Bjork calls <i>desirable difficulties<\/i>, i.e. making things difficult for yourself in a good way to optimise learning (Bjork &amp; Bjork, 2011).<\/p>\n<p class=\"western\"><b>Use an LLM for support, not substance<\/b><\/p>\n<p class=\"western\">Use LLMs for routine help, e.g. checking grammar, summarising prompts, generating examples, &amp; concept checking questions, but don\u2019t let it think for you. Core cognitive tasks like planning, drafting, &amp; revising should be your job.<\/p>\n<p class=\"western\"><b>Engage actively &amp; critically<\/b><\/p>\n<p class=\"western\">If you do use an LLM, analyse its output. Ask yourself: \u201cDoes this make sense? Could I say it better? What\u2019s missing?\u201d This kind of active involvement strengthens memory, understanding, &amp; critical thinking.<\/p>\n<p class=\"western\"><b>Keep your writing yours<\/b><\/p>\n<p class=\"western\">Avoid sounding like everyone else. Be original. Think critically. Use your own experiences &amp; voice. An LLM can make helpful suggestions but you bring your understanding of the content &amp; connect the concepts.<\/p>\n<h3 class=\"western\">In conclusion<\/h3>\n<p class=\"western\">LLMs like ChatGPT are useful assistants but they\u2019re not substitutes for your own mind. If you rely on them too much, you risk weakening the very skills you&#8217;re in school, college, or university to develop: memory, creativity, reasoning, &amp; independence. Use them but don\u2019t let them do the thinking for you. Durable learning is the result of thinking hard for sustained periods of time. Don\u2019t miss those opportunities to strengthen your brain!<\/p>\n<h3 class=\"western\">References<\/h3>\n<ul>\n<li>Bjork, R., &amp; Bjork, E. L. (2011). Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning. In Psychology and the Real World (pp. 55\u201364). Worth Publishers. <a href=\"https:\/\/bjorklab.psych.ucla.edu\/wp-content\/uploads\/sites\/13\/2016\/04\/EBjork_RBjork_2011.pdf\" target=\"_blank\" rel=\"noopener\">https:\/\/bjorklab.psych.ucla.edu\/wp-content\/uploads\/sites\/13\/2016\/04\/EBjork_RBjork_2011.pdf<\/a><\/li>\n<li>Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., &amp; Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (No. arXiv:2506.08872). arXiv. <a href=\"https:\/\/doi.org\/10.48550\/arXiv.2506.08872\" target=\"_blank\" rel=\"noopener\">https:\/\/doi.org\/10.48550\/arXiv.2506.08872<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Haga clic aqu\u00ed para la traducci\u00f3n al espa\u00f1ol (Google Translate) The following is my response to the current discourse surrounding a recent pre-publication paper, \u201cYour Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task\u201d (Kosmyna et al., 2025) from a study that examined &amp; compared the effects of&hellip; <a class=\"more-link\" href=\"https:\/\/matbury.com\/wordpress\/index.php\/2025\/07\/16\/think-before-you-rely-on-chatgpt-advice-for-students-using-an-llm-to-study\/\">Continue reading <span class=\"screen-reader-text\">Think Before You Rely on ChatGPT: Advice for students using an LLM to study<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":4871,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[122,7,115,116,121],"tags":[],"_links":{"self":[{"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/posts\/4870"}],"collection":[{"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/comments?post=4870"}],"version-history":[{"count":2,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/posts\/4870\/revisions"}],"predecessor-version":[{"id":4929,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/posts\/4870\/revisions\/4929"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/media\/4871"}],"wp:attachment":[{"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/media?parent=4870"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/categories?post=4870"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/matbury.com\/wordpress\/index.php\/wp-json\/wp\/v2\/tags?post=4870"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}