May 19, 2025 • 25 mins
The Cognitive Cost of AI: A Deep Dive into Learning and Development
In this episode of PsyberSpace, host Leslie Poston explores how artificial intelligence is reshaping cognition, particularly in education. The discussion focuses on the developmental impact of AI on children and the cognitive trade-offs associated with relying on AI tools. The episode delves into the science behind why traditional methods like handwriting are essential for learning, memory, and cognitive development. It also highlights the risks of cognitive offloading and suggests strategies for integrating AI thoughtfully without compromising critical thinking and problem-solving skills. The takeaway: while AI offers many conveniences, it’s crucial to balance its use to foster resilient, capable minds.
00:00 Introduction to AI and Cognitive Impact
01:01 AI in Education: Potential and Pitfalls
01:40 Cognitive Offloading and Dependency
04:30 The Importance of Handwriting
07:47 Developmental Concerns with Early AI Use
09:38 Critical Thinking and AI Reliance
11:25 Guardrails and Solutions
13:04 The Role of Productive Struggle
16:17 AI Hallucinations and Misinformation
18:43 Strategies for Thoughtful AI Integration
24:19 Conclusion and Call to Action
VOTE FOR PSYBERSPACE THIS MONTH! I'm nominated for a Women in Podcasting award! Voting is open during the month of May - I'd love your support. Note: You don't have to vote for every category, just look for PsyberSpace in Science Podcasts at this link https://www.womenpodcasters.com/awards-voting?sc=27127076032d7fd70ddcab921ce8652fb78353d03
Research
Armitage, K. L., Naulls, S., & Gillett, H. R. (2023). Cognitive offloading of memory for visual pattern learning: Benefits for performance but costs for long-term learning. Cognitive Research: Principles and Implications, 8(1), 60.
Berninger, V. W., Abbott, R. D., Augsburger, A., & Garcia, N. (2009). Comparison of pen and keyboard transcription modes in children with and without learning disabilities. Learning Disability Quarterly, 32(3), 123-141.
Bjork, E. L., & Bjork, R. A. (2011). Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning. In M. A. Gernsbacher, R. W. Pew, L. M. Hough, & J. R. Pomerantz (Eds.), Psychology and the real world: Essays illustrating fundamental contributions to society (pp. 56-64). Worth Publishers.
Boaler, J. (2016). Mathematical mindsets: Unleashing students' potential through creative math, inspiring messages and innovative teaching. Jossey-Bass.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11(6), 671-684.
Dunn, T. L., & Risko, E. F. (2016). Toward a metacognitive account of cognitive offloading. Cognitive Science, 40(5), 1080-1127.
Dweck, C. S. (2006). Mindset: The new psychology of success. Random House.
Eskritt, M., & Ma, S. (2014). Intentional forgetting: Note-taking as a naturalistic example. Memory & Cognition, 42(2), 237-246.
Fiorella, L., & Mayer, R. E. (2016). Eight ways to promote generative learning. Educational Psychology Review, 28(4), 717-741.
Gerlich, M. (2023). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6.
Hatano, A., Ogawa, K., Yamashita, M., & Yamamoto, H. (2021). Active memory processing mediated by hippocampal theta oscillations is important for memory stabilization during wake and sleep. Scientific Reports, 11, 12135.
Ihara, A. S., Nakajima, K., Kake, A., Ishimaru, K., Osugi, K., & Naruse, Y. (2021). Advantage of handwriting over typing on learning words: Evidence from an N4
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Leslie Poston (00:11):
Welcome back to PsyberSpace. I'm your host,
Leslie Posten. This week, we'reexamining how artificial
intelligence is reshaping ourbrains and not necessarily for
the better. We're diving intothe growing body of research
showing that while AI might makelife easier, it's also changing
how we think, how we remember,how we learn. Specifically,
(00:34):
we're going to focus on thedevelopmental impacts,
especially on children, and thecognitive trade offs of relying
on tools that do our thinkingfor us.
We'll explore the science on whyhandwriting notes, struggling to
learn new things, and buildingmemory the hard way are still
essential. Spoiler alert (00:51):
the harder path is usually the
better one for your brain. Let'sget into it. In the past few
years, large language modelgenerative AI tools like
ChadGPT, Conmigo, and Grammarlyhave found their way into
classrooms at lightning speed.I'll be referring to large
(01:15):
language model generative AItools simply as AI for the rest
of the show.
On paper, the potential for AIand education seems incredible.
Personalized learning, twentyfour seven homework help,
adaptive tutoring systems. Butwhen we look under the hood, the
picture isn't all rosy. We'restarting to see some troubling
(01:36):
patterns, especially amongyounger students. One of the key
issues is what's known ascognitive offloading.
That means we outsourcethinking, memory, or decision
making to an external tool, inthis case, AI. It's similar to
using GPS so much that youforget how to navigate your own
(01:57):
neighborhood. In classrooms, AIis often used to provide answers
without requiring the student tostruggle through a problem. This
is happening even as high ascollege institutions at this
point, so it's not isolated tochildren. The result?
Students learn less deeply,remember less, and retain fewer
(02:17):
problem solving skills. Researchhas found a significant negative
correlation between frequent AItool use and critical thinking
abilities. One study showed thatparticipants who heavily relied
on AI tools performed worse oncritical thinking assessments
compared to those who used thetools less frequently. This
(02:38):
effect was particularlypronounced among younger
individuals, while those withhigher education levels tended
to retain stronger criticalthinking skills regardless of AI
tool usage. What's happeninghere is a form of cognitive
dependency.
Our brains are remarkablyadaptive organs. They're
designed to conserve energy whenpossible. When we repeatedly
(03:00):
outsource mental tasks to AI,our brains essentially say,
great, I don't need to maintainthose neural pathways anymore.
It's similar to how a muscleatrophies when it's not used
regularly. The cognitive musclesresponsible for critical
analysis, deep reading, andcomplex problem solving begin to
weaken when AI does that heavylifting for us.
(03:23):
The more these tools becomefixtures in the learning
process, the more they riskbecoming cognitive crutches.
Teachers report students usingAI to bypass assignments
entirely. A student may useChatGPT to write an essay or
solve a math problem withoutunderstanding the content. The
cost isn't just academicintegrity. It's a decline in the
(03:46):
underlying thinking muscles thatwe rely on for life and for
society.
There's also a socio emotionalelement. Students rely on AI
tools early and often may neverdevelop the confidence that
comes from figuring somethingout on your own. That internal
sense of mastery matters. And ifwe give it away too early, it's
(04:07):
difficult to build back later.Psychologists have long
recognized the importance ofself efficacy, that belief in
your own ability to accomplishtasks and overcome challenges
and cognitive development.
When students consistently turnto AI at the first sign of
difficulty, they miss essentialopportunities to build that self
(04:28):
efficacy. Remember notebooks,pens and pencils? There is a
mountain of research showingthat handwriting is far better
for learning than typing ortapping. When we write by hand,
we engage motor functions,visual pathways, and memory
centers in the brainsimultaneously. This
(04:49):
multisensory process helps uslearn and retain more
information.
Multiple neuroimaging studieshave documented dramatic
differences in brain activationbetween handwriting and typing.
One study found that whenstudents wrote by hand,
researchers observed increasedconnectivity across visual
regions, sensory processingareas, and the motor cortex
(05:13):
brain patterns vital for memoryformation. Typing, on the other
hand, led to minimal activationin the same areas. This isn't
just about a nostalgia for penand paper. The science is clear.
Handwriting activatesspecialized brain circuits that
integrate physical movement,visual perception, and spatial
(05:33):
awareness. Specifically, ittriggers regions in the brain's
parietal and occipital lobesthat are less active during
typing. These regions arecritical for processing
information at a deeper leveland encoding it into long term
memory. It's what scientistscall embodied cognition, this
idea that physical experienceshapes cognitive processing.
(05:56):
Typing tends to be more passive.
When you type, especiallyquickly, you often transcribe
rather than process, and thatdistinction is so important.
Multiple studies have found thatstudents who type notes tend to
do worse on conceptual questionscompared to those who handwrite
them. Why? Because handwritingforces you to rephrase, to
(06:18):
summarize, and to synthesize inthe moment. You can't possibly
handwrite as fast as someonespeaks, so your brain has to
make decisions about what'simportant.
That extra cognitive effort,deciding what to write, creates
stronger memory traces. Consideranother study where researchers
compared note taking methodsamong college students. Students
(06:42):
who took handwritten notesshowed better conceptual
understanding and performedbetter on assessments taken a
week later compared to laptopnotetakers. The laptop users
took more notes verbatim whilethe handwriters were processing
and rephrasing information inreal time, which is a more
active form of learning. Nowwe're getting back to that sense
(07:04):
of embodied cognition that wetalked about a few minutes ago,
that physical act of writing,because it encodes the
information in a different andmore durable way.
Like we said before, it'sespecially important for younger
learners. Their brains are stillforming. They're forming those
neural pathways tied to finemotor skills and language
processing. Take awayhandwriting, and you take away a
(07:25):
developmental tool. And yetknowing all of this, many
schools are moving away fromhandwriting instruction
altogether, especially cursivewriting.
That's a mistake, plain andsimple. We know it supports
learning, and we know that itsupports memory. It might not be
as fast or shiny as AI, but it'smuch more effective in the long
run. Developmental psychologistsand neuroscientists are raising
(07:51):
red flags about the prematureuse of AI in classrooms. Recent
research out of Harvard andother institutions is showing us
that early use of AI tools canshort circuit the development of
core cognitive skills, attentionspan, working memory, and
executive function.
These are the building blocks oflearning. Here's where things
can get even more concerning.Kids' brains are especially
(08:14):
sensitive to how and what theylearn. Introducing AI into early
education may shape not justwhat they know, but how they
know. And once those habitsform, they're hard to unlearn.
The more a child relies on AI toanswer questions or complete
tasks, the less opportunity theyhave to practice curiosity,
exploration, and frustrationtolerance. These all sound like
(08:38):
soft skills, but they're not.They're essential parts of a
growing brain that can thinkcritically and independently.
Studies suggest that even youngchildren are quick to offload
cognitive processes to externaltools if given the opportunity.
This is worrying because it'shappening during critical
developmental periods when ourneural pathways are being
established.
(08:59):
The pattern of immediatelyturning to AI for answers rather
than engaging in hard mentaleffort may establish lifelong
habits that diminish theircognitive abilities. And it's
not just cognition that's atstake. There are also concerns
about how AI interactions shapesocial development. Children
engaging with assistance maydevelop different communication
(09:20):
habits, fewer interpersonalskills, and a reduced ability to
read emotional cues. In short,early exposure to AI isn't
neutral.
It shapes the brain. And unlesswe put some guardrails in place,
it might shape our kids' brainsin ways we come to regret. Let's
talk about what happens tothinking when AI fills in the
(09:42):
blanks for us. Across all agegroups, we're seeing a quiet
decline in critical thinkingskills. Students and adults for
that matter who rely on AI tooheavily are less likely to
analyze, question, or even todoubt the information that they
passively receive.
It's not just that AI is alwayswrong, though it can be. It's
(10:04):
that it flattens complexity.When an answer is generated
instantly and without effort, itrobs us of the chance to form
our own reasoning pathways.Those are the skills that help
us evaluate, argue, interpret,and innovate. Researchers have
found that frequent AI usersincreasingly display what's
called epistemic dependence.
(10:25):
They begin to doubt their ownreasoning, when confronted with
AI that sounds moreauthoritative than they do,
instead of asking, Is this true?They ask, why bother thinking
about it? The AI already knows.Teachers across the country are
reporting that students arehanding in work that looks
polished but lacks depth. It'sharder to spot because AI
(10:47):
grammar is good and theformatting is clean, but
underneath, it's hollow.
There's no depth. And many ofthe things that are being handed
in are the same. This shouldscare us. A generation of
students who can passassignments but can't form
arguments is a problem fordemocracy as well as education.
(11:07):
Even in higher education,professors are struggling with
how to balance the benefits ofAI tools with the need for
genuine intellectual engagement.
It's not an easy line to draw,but we have to draw it.
Otherwise, we're just trainingpeople to outsource their
thinking permanently. So what dowe do about all this? One answer
is clear (11:29):
we need stronger guardrails in place, especially
before college. K-twelveeducation, in my opinion, should
be focused on helping childrendevelop strong, resilient
brains.
That means limiting reliance ontools that give them answers too
easily. Some schools are alreadyexperimenting with this. They
allow limited AI use in highschool, but not before. Others
(11:52):
are creating AI free zones forcertain types of assignments.
Policymakers and educators mustwork together to develop ethical
frameworks for AI and education.
These need to be rooted incognitive science, not tech
industry hype. And they need toconsider developmental
appropriateness. Just because atool exists doesn't mean it
(12:13):
should be used at every age.Educators can implement what's
called productive struggle intotheir teaching approach. This
concept emphasizes the value ofletting students grapple with
challenging tasks that areslightly beyond their current
abilities.
Neuroscience research shows thatthis kind of effortful
engagement actually producesmyelin, a substance that
(12:35):
strengthens neural connectionsin the brain. Struggle, and the
right amount, literally buildsbetter brain circuitry. Parents
also have a role to play. Theycan advocate for policies that
protect cognitive development.They can choose learning tools
carefully and encourage kids todo things the hard way
sometimes.
It's good for them, even if it'sfrustrating in the moment.
(12:57):
Guardrails aren't anti tech.They're pro brain, and we need
them now. There's somethingpsychologically powerful about
struggling with a problem andthen figuring it out. That
moment of breakthrough wiresyour brain.
It creates resilience,physically creates resilience.
(13:18):
It builds what psychologistscall productive failure. AI
eliminates that struggle. Itgives us answers before we even
finish forming the question, andthat might feel good in the
moment, but it short circuitslearning and brain development.
And struggle isn't failure.
Struggle is the process by whichwe grow. In fact, studies show
(13:38):
that people remember thingsbetter when they've had to work
harder to understand them. Thatfriction matters. It's how long
term memory is formed. Let'stalk about myelin again.
As we mentioned before,educational neuroscience
research reveals that when wework through difficult problems,
our brains produce myelin, asubstance that coats neural
(13:58):
pathways and increases thestrength and efficiency of brain
signals. The myelination processis critical for building faster
and more efficient neuralpathways, essentially creating
better roads in the brain forinformation to travel. Easy
answers from AI skip thisnecessary brain building
process. The concept ofdesirable difficulties in
(14:19):
learning has been wellestablished in cognitive
psychology. Coined byresearchers Robert and Elizabeth
Bjork, this theory suggests thatintroducing certain difficulties
into the learning process,difficulties that are relevant
to the material and appropriateto the learner's ability level,
can significantly enhance longterm retention and transfer of
(14:40):
knowledge.
These beneficial challengesforce the brain to engage more
deeply with the material,forming stronger neural
connections in the process.Carol Dweck's research on growth
mindset further supports thisidea. Her work shows that
individuals who embracechallenges and persist through
difficulties develop greatercognitive resilience and
(15:02):
ultimately achieve more. Whenstudents understand that
struggle is not a sign ofinability but rather a pathway
to growth, they're more likelyto engage with difficult
material and develop deeperunderstanding. Consider a
classic experiment wherestudents were divided into two
groups learning the samematerial.
One group studied underconditions that made learning
(15:24):
easy, while the other facedconditions that required more
effort. In immediate testing,the easy group performed better.
But when tested weeks later, thegroup that had to struggle
showed far superior retention.This pattern has been replicated
across various age groups andacross various subject areas in
several research studies,confirming that shortcuts can
(15:46):
often lead to shorter lastinglearning. When students rely on
AI for these quick answers,they're missing out on the
benefits of that friction.
They become more passive. Theylearn less deeply. They might
even lose confidence in theirown problem solving abilities,
thinking they can't do itwithout AI. This is a cultural
shift and one we need toreverse. We must reframe
(16:08):
struggle as strength, not aweakness, and we need to make
space for it in our classrooms,our homes, and our workplaces.
One of the lesser understood butincreasingly dangerous risks of
widespread AI use is thephenomenon known as
hallucination, which is when anAI confidently produces
information that is factuallycorrect, that is factually
(16:31):
incorrect, biased, or entirelyfabricated. These hallucinations
aren't glitches in thetraditional sense. They're a
feature of how large languagemodels work. They generate
plausible sounding responses bypredicting patterns in language,
not by checking truth. Thatmeans the better the model is at
sounding authoritative, theharder it can be to spot one is
(16:52):
wrong.
Now combine that with thedecline in human critical
thinking and cognitive effortthat we've been discussing. When
students and even adults getused to trusting AI responses
without verification, they losethe muscle memory required to
fact check, cross reference, andquestion. This is especially
troubling in a world alreadyflooded with misinformation. If
(17:15):
our tools are unreliable and ourability to challenge them is
weakened, that is a recipe formass confusion and manipulation.
There have already been realworld consequences.
AI has been caught fabricatinglegal citations, misrepresenting
research studies, and eveninventing quotes from public
figures. In academic,government, and professional
(17:37):
settings, this has led toembarrassing and sometimes
costly errors. But in K through12 education, it's even more
insidious. Young learners maynot yet have the knowledge base
or skepticism to question whatan AI tells them, especially
when it sounds confident andfluent. There's also a
psychological effect.
(17:58):
When humans repeatedly outsourcethinking to a machine that
sounds smarter or moreauthoritative than they do, they
start to doubt their ownreasoning. This is where learned
helplessness and epistemicdependence take hold. Like we
said before, remember, insteadof asking, is this true? They're
asking, bother thinking about itif the AI already knows?
(18:20):
Rebuilding that internalcompass, what researchers call
epistemic vigilance, iscritical.
We need to teach people not justhow to use AI, but how to doubt
it, how to challenge it, and howto confirm its outputs through
their own mental effort. Withoutthat, we're risking creating a
future where hallucinationsaren't just an AI problem,
they're a human one. Now, thisdoesn't mean we should ban AI
(18:45):
from education entirely. That'sneither realistic nor helpful.
What we need is thoughtfulintegration.
Used well, AI can be a tool forexpanding access, enhancing
feedback, and personalizinginstruction. But we must use it
intentionally. That meansdesigning learning experiences
that prioritize cognition, notconvenience. For example, AI
(19:09):
could be used to give studentsfeedback on drafts, but not to
write the drafts. It couldsuggest questions to explore,
but not answer them outright.
Educators can set clearboundaries. They can create
assignments that requirepersonal insight, reflection, or
synthesis, things AI stillstruggles to fake. And they can
teach students how to use AI asa thinking partner, not a
(19:32):
shortcut. This research doesn'tsuggest that all AI use is
harmful. In fact, moderate andstrategic use of AI tools can
enhance learning when they'reused to scaffold more complex
cognitive tasks.
The key is ensuring that AIserves as an enhancement to
human thinking rather than areplacement for it. This means
explicitly teachingmetacognitive skills alongside
(19:54):
AI use, helping studentsunderstand when to use AI and
when to rely on their owncognitive resources. When AI is
used to provoke thought insteadof replace it, it becomes a tool
for learning. And that's thesweet spot. That's what we
should aim for.
If you're a parent, a teacher,or a policymaker, here are some
(20:16):
steps you can take to protectcognitive development in the age
of AI. First, limit AI tool usebefore high school. The younger
the brain, the more it needs toengage with the world through
real effort and real feedback.Create opportunities for
productive struggle wherechildren learn to persist
through challenges withoutimmediately turning to AI for
(20:37):
solutions. For elementary andmiddle school students, consider
designated tech free learningperiods where students must rely
on their internal resources tosolve problems.
This doesn't mean abandoningtechnology entirely, but rather
being strategic about how andwhen it's used. Second,
reintroduce handwriting whereverpossible. Encourage students to
(21:00):
write notes, journal, orbrainstorm on paper. The
benefits are well established.Multiple studies have confirmed
that handwriting activatescritical brain connections
across the visual, motor, andmemory regions in ways that
typing simply can't replicate.
So even for older students whoprefer typing for longer
assignments, encouraginghandwritten outlines or initial
(21:21):
drafts to engage these neuralbenefits during the conceptual
phase of work is beneficial. Andit can be done even in the age
of remote learning and remotework by holding writing sessions
on Zoom or in the classroomvideo software, then having the
students scan in the work thatthey've done in real time.
Third, design learningexperiences that build in
(21:42):
struggle. Let students wrestlewith hard questions. Let them
fail safely.
Let them grow. Research clearlyshows that this kind of
productive struggle is not justabout building character. It's
physically strengthening neuralpathways through the production
of myelin in the brain. Createwhat education researchers call
low floor, high ceiling tasks: activities that are accessible (22:00):
undefined
to all students, but that can beextended to challenge even the
most advanced learners. Createwhat education researchers call
low floor, high ceiling tasksactivities that are accessible
to all students that can beextended to challenge even the
most advanced learners.
(22:21):
This approach ensures everyoneexperiences an appropriate level
of productive struggle. Fourth,talk openly about AI. Make it
clear that it's a tool and not acrutch. Teach digital literacy
and critical thinking alongsideany tech instruction.
Specifically, teach studentsabout AI hallucinations and how
(22:41):
to verify AI generatedinformation as this builds
essential fact checking habitsthat strengthen their own
cognitive independence.
This is especially important aswe get more sophisticated video
AI tools as well. Developclassroom protocols for
appropriate AI use. Forinstance, students might use AI
(23:01):
to help brainstorm ideas orcheck their work, but not to
generate complete assignments.These guardrails help students
develop discernment about whenAI enhances learning versus when
it undermines it. Fifth, modelwhat deep thinking looks like.
Whether you're a teacher or aparent, your approach to problem
solving sets the tone. Show yourkids that thinking hard is worth
(23:24):
it. Demonstrate the value ofslow, deliberate thought in an
age of instant answers. Thinkout loud as you work through
problems, showing students thatconfusion and false starts are
normal parts of the thinkingprocess. When you encounter
information online, model how toevaluate its credibility rather
than accepting it at face value.
(23:46):
Finally, advocate for balancedtechnology policies in schools.
Get involved in curriculumdiscussions at your school. Ask
questions about how technologyis being integrated and what
safeguards are in place toensure it's enhancing rather
than replacing criticalthinking. Share research on
cognitive development withadministrators and fellow
parents. Many schools are stillfiguring out how to navigate AI
(24:09):
tools in education, and informedparent voices can help shape
thoughtful policies that protectcognitive development while
embracing beneficial innovation.
And remember, we're not arguingagainst technology. We're
advocating for its thoughtfulintegration in ways that enhance
rather than diminish humancognitive capacities. The goal
(24:30):
is not to hold back progress,but to ensure that in our rush
to embrace AI's convenience, wedon't accidentally outsource the
very mental processes that makeus human. Thanks for listening
to PsyberSpace. I'm your host,Leslie Posten, signing off.
This episode explored thecomplicated relationship between
AI and learning, especially howit's reshaping cognition for the
(24:51):
next generation. The science isclear. The brain needs effort,
friction, handwriting, andreflection to grow strong. If we
hand over all the heavy liftingto AI, we might find that we've
traded ease for erosion. A bitof housekeeping.
We were nominated for a Women inPodcasting Award for the second
year in a row. We're so excitedabout it. Thank you so much. The
(25:14):
link to vote for the show willbe in the show notes, and I
would love your vote for BestScience Podcast. As always,
until next time, stay curious,and don't forget to subscribe so
you never miss a week.
Welcome back to PsyberSpace. I'm your host,
Leslie Posten. This week, we'reexamining how artificial
intelligence is reshaping ourbrains and not necessarily for
the better. We're diving intothe growing body of research
showing that while AI might makelife easier, it's also changing
how we think, how we remember,how we learn. Specifically,
(00:34):
we're going to focus on thedevelopmental impacts,
especially on children, and thecognitive trade offs of relying
on tools that do our thinkingfor us.
We'll explore the science on whyhandwriting notes, struggling to
learn new things, and buildingmemory the hard way are still
essential. Spoiler alert (00:51):
the harder path is usually the
better one for your brain. Let'sget into it. In the past few
years, large language modelgenerative AI tools like
ChadGPT, Conmigo, and Grammarlyhave found their way into
classrooms at lightning speed.I'll be referring to large
(01:15):
language model generative AItools simply as AI for the rest
of the show.
On paper, the potential for AIand education seems incredible.
Personalized learning, twentyfour seven homework help,
adaptive tutoring systems. Butwhen we look under the hood, the
picture isn't all rosy. We'restarting to see some troubling
(01:36):
patterns, especially amongyounger students. One of the key
issues is what's known ascognitive offloading.
That means we outsourcethinking, memory, or decision
making to an external tool, inthis case, AI. It's similar to
using GPS so much that youforget how to navigate your own
(01:57):
neighborhood. In classrooms, AIis often used to provide answers
without requiring the student tostruggle through a problem. This
is happening even as high ascollege institutions at this
point, so it's not isolated tochildren. The result?
Students learn less deeply,remember less, and retain fewer
(02:17):
problem solving skills. Researchhas found a significant negative
correlation between frequent AItool use and critical thinking
abilities. One study showed thatparticipants who heavily relied
on AI tools performed worse oncritical thinking assessments
compared to those who used thetools less frequently. This
(02:38):
effect was particularlypronounced among younger
individuals, while those withhigher education levels tended
to retain stronger criticalthinking skills regardless of AI
tool usage. What's happeninghere is a form of cognitive
dependency.
Our brains are remarkablyadaptive organs. They're
designed to conserve energy whenpossible. When we repeatedly
(03:00):
outsource mental tasks to AI,our brains essentially say,
great, I don't need to maintainthose neural pathways anymore.
It's similar to how a muscleatrophies when it's not used
regularly. The cognitive musclesresponsible for critical
analysis, deep reading, andcomplex problem solving begin to
weaken when AI does that heavylifting for us.
(03:23):
The more these tools becomefixtures in the learning
process, the more they riskbecoming cognitive crutches.
Teachers report students usingAI to bypass assignments
entirely. A student may useChatGPT to write an essay or
solve a math problem withoutunderstanding the content. The
cost isn't just academicintegrity. It's a decline in the
(03:46):
underlying thinking muscles thatwe rely on for life and for
society.
There's also a socio emotionalelement. Students rely on AI
tools early and often may neverdevelop the confidence that
comes from figuring somethingout on your own. That internal
sense of mastery matters. And ifwe give it away too early, it's
(04:07):
difficult to build back later.Psychologists have long
recognized the importance ofself efficacy, that belief in
your own ability to accomplishtasks and overcome challenges
and cognitive development.
When students consistently turnto AI at the first sign of
difficulty, they miss essentialopportunities to build that self
(04:28):
efficacy. Remember notebooks,pens and pencils? There is a
mountain of research showingthat handwriting is far better
for learning than typing ortapping. When we write by hand,
we engage motor functions,visual pathways, and memory
centers in the brainsimultaneously. This
(04:49):
multisensory process helps uslearn and retain more
information.
Multiple neuroimaging studieshave documented dramatic
differences in brain activationbetween handwriting and typing.
One study found that whenstudents wrote by hand,
researchers observed increasedconnectivity across visual
regions, sensory processingareas, and the motor cortex
(05:13):
brain patterns vital for memoryformation. Typing, on the other
hand, led to minimal activationin the same areas. This isn't
just about a nostalgia for penand paper. The science is clear.
Handwriting activatesspecialized brain circuits that
integrate physical movement,visual perception, and spatial
(05:33):
awareness. Specifically, ittriggers regions in the brain's
parietal and occipital lobesthat are less active during
typing. These regions arecritical for processing
information at a deeper leveland encoding it into long term
memory. It's what scientistscall embodied cognition, this
idea that physical experienceshapes cognitive processing.
(05:56):
Typing tends to be more passive.
When you type, especiallyquickly, you often transcribe
rather than process, and thatdistinction is so important.
Multiple studies have found thatstudents who type notes tend to
do worse on conceptual questionscompared to those who handwrite
them. Why? Because handwritingforces you to rephrase, to
(06:18):
summarize, and to synthesize inthe moment. You can't possibly
handwrite as fast as someonespeaks, so your brain has to
make decisions about what'simportant.
That extra cognitive effort,deciding what to write, creates
stronger memory traces. Consideranother study where researchers
compared note taking methodsamong college students. Students
(06:42):
who took handwritten notesshowed better conceptual
understanding and performedbetter on assessments taken a
week later compared to laptopnotetakers. The laptop users
took more notes verbatim whilethe handwriters were processing
and rephrasing information inreal time, which is a more
active form of learning. Nowwe're getting back to that sense
(07:04):
of embodied cognition that wetalked about a few minutes ago,
that physical act of writing,because it encodes the
information in a different andmore durable way.
Like we said before, it'sespecially important for younger
learners. Their brains are stillforming. They're forming those
neural pathways tied to finemotor skills and language
processing. Take awayhandwriting, and you take away a
(07:25):
developmental tool. And yetknowing all of this, many
schools are moving away fromhandwriting instruction
altogether, especially cursivewriting.
That's a mistake, plain andsimple. We know it supports
learning, and we know that itsupports memory. It might not be
as fast or shiny as AI, but it'smuch more effective in the long
run. Developmental psychologistsand neuroscientists are raising
(07:51):
red flags about the prematureuse of AI in classrooms. Recent
research out of Harvard andother institutions is showing us
that early use of AI tools canshort circuit the development of
core cognitive skills, attentionspan, working memory, and
executive function.
These are the building blocks oflearning. Here's where things
can get even more concerning.Kids' brains are especially
(08:14):
sensitive to how and what theylearn. Introducing AI into early
education may shape not justwhat they know, but how they
know. And once those habitsform, they're hard to unlearn.
The more a child relies on AI toanswer questions or complete
tasks, the less opportunity theyhave to practice curiosity,
exploration, and frustrationtolerance. These all sound like
(08:38):
soft skills, but they're not.They're essential parts of a
growing brain that can thinkcritically and independently.
Studies suggest that even youngchildren are quick to offload
cognitive processes to externaltools if given the opportunity.
This is worrying because it'shappening during critical
developmental periods when ourneural pathways are being
established.
(08:59):
The pattern of immediatelyturning to AI for answers rather
than engaging in hard mentaleffort may establish lifelong
habits that diminish theircognitive abilities. And it's
not just cognition that's atstake. There are also concerns
about how AI interactions shapesocial development. Children
engaging with assistance maydevelop different communication
(09:20):
habits, fewer interpersonalskills, and a reduced ability to
read emotional cues. In short,early exposure to AI isn't
neutral.
It shapes the brain. And unlesswe put some guardrails in place,
it might shape our kids' brainsin ways we come to regret. Let's
talk about what happens tothinking when AI fills in the
(09:42):
blanks for us. Across all agegroups, we're seeing a quiet
decline in critical thinkingskills. Students and adults for
that matter who rely on AI tooheavily are less likely to
analyze, question, or even todoubt the information that they
passively receive.
It's not just that AI is alwayswrong, though it can be. It's
(10:04):
that it flattens complexity.When an answer is generated
instantly and without effort, itrobs us of the chance to form
our own reasoning pathways.Those are the skills that help
us evaluate, argue, interpret,and innovate. Researchers have
found that frequent AI usersincreasingly display what's
called epistemic dependence.
(10:25):
They begin to doubt their ownreasoning, when confronted with
AI that sounds moreauthoritative than they do,
instead of asking, Is this true?They ask, why bother thinking
about it? The AI already knows.Teachers across the country are
reporting that students arehanding in work that looks
polished but lacks depth. It'sharder to spot because AI
(10:47):
grammar is good and theformatting is clean, but
underneath, it's hollow.
There's no depth. And many ofthe things that are being handed
in are the same. This shouldscare us. A generation of
students who can passassignments but can't form
arguments is a problem fordemocracy as well as education.
(11:07):
Even in higher education,professors are struggling with
how to balance the benefits ofAI tools with the need for
genuine intellectual engagement.
It's not an easy line to draw,but we have to draw it.
Otherwise, we're just trainingpeople to outsource their
thinking permanently. So what dowe do about all this? One answer
is clear (11:29):
we need stronger guardrails in place, especially
before college. K-twelveeducation, in my opinion, should
be focused on helping childrendevelop strong, resilient
brains.
That means limiting reliance ontools that give them answers too
easily. Some schools are alreadyexperimenting with this. They
allow limited AI use in highschool, but not before. Others
(11:52):
are creating AI free zones forcertain types of assignments.
Policymakers and educators mustwork together to develop ethical
frameworks for AI and education.
These need to be rooted incognitive science, not tech
industry hype. And they need toconsider developmental
appropriateness. Just because atool exists doesn't mean it
(12:13):
should be used at every age.Educators can implement what's
called productive struggle intotheir teaching approach. This
concept emphasizes the value ofletting students grapple with
challenging tasks that areslightly beyond their current
abilities.
Neuroscience research shows thatthis kind of effortful
engagement actually producesmyelin, a substance that
(12:35):
strengthens neural connectionsin the brain. Struggle, and the
right amount, literally buildsbetter brain circuitry. Parents
also have a role to play. Theycan advocate for policies that
protect cognitive development.They can choose learning tools
carefully and encourage kids todo things the hard way
sometimes.
It's good for them, even if it'sfrustrating in the moment.
(12:57):
Guardrails aren't anti tech.They're pro brain, and we need
them now. There's somethingpsychologically powerful about
struggling with a problem andthen figuring it out. That
moment of breakthrough wiresyour brain.
It creates resilience,physically creates resilience.
(13:18):
It builds what psychologistscall productive failure. AI
eliminates that struggle. Itgives us answers before we even
finish forming the question, andthat might feel good in the
moment, but it short circuitslearning and brain development.
And struggle isn't failure.
Struggle is the process by whichwe grow. In fact, studies show
(13:38):
that people remember thingsbetter when they've had to work
harder to understand them. Thatfriction matters. It's how long
term memory is formed. Let'stalk about myelin again.
As we mentioned before,educational neuroscience
research reveals that when wework through difficult problems,
our brains produce myelin, asubstance that coats neural
(13:58):
pathways and increases thestrength and efficiency of brain
signals. The myelination processis critical for building faster
and more efficient neuralpathways, essentially creating
better roads in the brain forinformation to travel. Easy
answers from AI skip thisnecessary brain building
process. The concept ofdesirable difficulties in
(14:19):
learning has been wellestablished in cognitive
psychology. Coined byresearchers Robert and Elizabeth
Bjork, this theory suggests thatintroducing certain difficulties
into the learning process,difficulties that are relevant
to the material and appropriateto the learner's ability level,
can significantly enhance longterm retention and transfer of
(14:40):
knowledge.
These beneficial challengesforce the brain to engage more
deeply with the material,forming stronger neural
connections in the process.Carol Dweck's research on growth
mindset further supports thisidea. Her work shows that
individuals who embracechallenges and persist through
difficulties develop greatercognitive resilience and
(15:02):
ultimately achieve more. Whenstudents understand that
struggle is not a sign ofinability but rather a pathway
to growth, they're more likelyto engage with difficult
material and develop deeperunderstanding. Consider a
classic experiment wherestudents were divided into two
groups learning the samematerial.
One group studied underconditions that made learning
(15:24):
easy, while the other facedconditions that required more
effort. In immediate testing,the easy group performed better.
But when tested weeks later, thegroup that had to struggle
showed far superior retention.This pattern has been replicated
across various age groups andacross various subject areas in
several research studies,confirming that shortcuts can
(15:46):
often lead to shorter lastinglearning. When students rely on
AI for these quick answers,they're missing out on the
benefits of that friction.
They become more passive. Theylearn less deeply. They might
even lose confidence in theirown problem solving abilities,
thinking they can't do itwithout AI. This is a cultural
shift and one we need toreverse. We must reframe
(16:08):
struggle as strength, not aweakness, and we need to make
space for it in our classrooms,our homes, and our workplaces.
One of the lesser understood butincreasingly dangerous risks of
widespread AI use is thephenomenon known as
hallucination, which is when anAI confidently produces
information that is factuallycorrect, that is factually
(16:31):
incorrect, biased, or entirelyfabricated. These hallucinations
aren't glitches in thetraditional sense. They're a
feature of how large languagemodels work. They generate
plausible sounding responses bypredicting patterns in language,
not by checking truth. Thatmeans the better the model is at
sounding authoritative, theharder it can be to spot one is
(16:52):
wrong.
Now combine that with thedecline in human critical
thinking and cognitive effortthat we've been discussing. When
students and even adults getused to trusting AI responses
without verification, they losethe muscle memory required to
fact check, cross reference, andquestion. This is especially
troubling in a world alreadyflooded with misinformation. If
(17:15):
our tools are unreliable and ourability to challenge them is
weakened, that is a recipe formass confusion and manipulation.
There have already been realworld consequences.
AI has been caught fabricatinglegal citations, misrepresenting
research studies, and eveninventing quotes from public
figures. In academic,government, and professional
(17:37):
settings, this has led toembarrassing and sometimes
costly errors. But in K through12 education, it's even more
insidious. Young learners maynot yet have the knowledge base
or skepticism to question whatan AI tells them, especially
when it sounds confident andfluent. There's also a
psychological effect.
(17:58):
When humans repeatedly outsourcethinking to a machine that
sounds smarter or moreauthoritative than they do, they
start to doubt their ownreasoning. This is where learned
helplessness and epistemicdependence take hold. Like we
said before, remember, insteadof asking, is this true? They're
asking, bother thinking about itif the AI already knows?
(18:20):
Rebuilding that internalcompass, what researchers call
epistemic vigilance, iscritical.
We need to teach people not justhow to use AI, but how to doubt
it, how to challenge it, and howto confirm its outputs through
their own mental effort. Withoutthat, we're risking creating a
future where hallucinationsaren't just an AI problem,
they're a human one. Now, thisdoesn't mean we should ban AI
(18:45):
from education entirely. That'sneither realistic nor helpful.
What we need is thoughtfulintegration.
Used well, AI can be a tool forexpanding access, enhancing
feedback, and personalizinginstruction. But we must use it
intentionally. That meansdesigning learning experiences
that prioritize cognition, notconvenience. For example, AI
(19:09):
could be used to give studentsfeedback on drafts, but not to
write the drafts. It couldsuggest questions to explore,
but not answer them outright.
Educators can set clearboundaries. They can create
assignments that requirepersonal insight, reflection, or
synthesis, things AI stillstruggles to fake. And they can
teach students how to use AI asa thinking partner, not a
(19:32):
shortcut. This research doesn'tsuggest that all AI use is
harmful. In fact, moderate andstrategic use of AI tools can
enhance learning when they'reused to scaffold more complex
cognitive tasks.
The key is ensuring that AIserves as an enhancement to
human thinking rather than areplacement for it. This means
explicitly teachingmetacognitive skills alongside
(19:54):
AI use, helping studentsunderstand when to use AI and
when to rely on their owncognitive resources. When AI is
used to provoke thought insteadof replace it, it becomes a tool
for learning. And that's thesweet spot. That's what we
should aim for.
If you're a parent, a teacher,or a policymaker, here are some
(20:16):
steps you can take to protectcognitive development in the age
of AI. First, limit AI tool usebefore high school. The younger
the brain, the more it needs toengage with the world through
real effort and real feedback.Create opportunities for
productive struggle wherechildren learn to persist
through challenges withoutimmediately turning to AI for
(20:37):
solutions. For elementary andmiddle school students, consider
designated tech free learningperiods where students must rely
on their internal resources tosolve problems.
This doesn't mean abandoningtechnology entirely, but rather
being strategic about how andwhen it's used. Second,
reintroduce handwriting whereverpossible. Encourage students to
(21:00):
write notes, journal, orbrainstorm on paper. The
benefits are well established.Multiple studies have confirmed
that handwriting activatescritical brain connections
across the visual, motor, andmemory regions in ways that
typing simply can't replicate.
So even for older students whoprefer typing for longer
assignments, encouraginghandwritten outlines or initial
(21:21):
drafts to engage these neuralbenefits during the conceptual
phase of work is beneficial. Andit can be done even in the age
of remote learning and remotework by holding writing sessions
on Zoom or in the classroomvideo software, then having the
students scan in the work thatthey've done in real time.
Third, design learningexperiences that build in
(21:42):
struggle. Let students wrestlewith hard questions. Let them
fail safely.
Let them grow. Research clearlyshows that this kind of
productive struggle is not justabout building character. It's
physically strengthening neuralpathways through the production
of myelin in the brain. Createwhat education researchers call
low floor, high ceiling tasks: activities that are accessible (22:00):
undefined
to all students, but that can beextended to challenge even the
most advanced learners. Createwhat education researchers call
low floor, high ceiling tasksactivities that are accessible
to all students that can beextended to challenge even the
most advanced learners.
(22:21):
This approach ensures everyoneexperiences an appropriate level
of productive struggle. Fourth,talk openly about AI. Make it
clear that it's a tool and not acrutch. Teach digital literacy
and critical thinking alongsideany tech instruction.
Specifically, teach studentsabout AI hallucinations and how
(22:41):
to verify AI generatedinformation as this builds
essential fact checking habitsthat strengthen their own
cognitive independence.
This is especially important aswe get more sophisticated video
AI tools as well. Developclassroom protocols for
appropriate AI use. Forinstance, students might use AI
(23:01):
to help brainstorm ideas orcheck their work, but not to
generate complete assignments.These guardrails help students
develop discernment about whenAI enhances learning versus when
it undermines it. Fifth, modelwhat deep thinking looks like.
Whether you're a teacher or aparent, your approach to problem
solving sets the tone. Show yourkids that thinking hard is worth
(23:24):
it. Demonstrate the value ofslow, deliberate thought in an
age of instant answers. Thinkout loud as you work through
problems, showing students thatconfusion and false starts are
normal parts of the thinkingprocess. When you encounter
information online, model how toevaluate its credibility rather
than accepting it at face value.
(23:46):
Finally, advocate for balancedtechnology policies in schools.
Get involved in curriculumdiscussions at your school. Ask
questions about how technologyis being integrated and what
safeguards are in place toensure it's enhancing rather
than replacing criticalthinking. Share research on
cognitive development withadministrators and fellow
parents. Many schools are stillfiguring out how to navigate AI
(24:09):
tools in education, and informedparent voices can help shape
thoughtful policies that protectcognitive development while
embracing beneficial innovation.
And remember, we're not arguingagainst technology. We're
advocating for its thoughtfulintegration in ways that enhance
rather than diminish humancognitive capacities. The goal
(24:30):
is not to hold back progress,but to ensure that in our rush
to embrace AI's convenience, wedon't accidentally outsource the
very mental processes that makeus human. Thanks for listening
to PsyberSpace. I'm your host,Leslie Posten, signing off.
This episode explored thecomplicated relationship between
AI and learning, especially howit's reshaping cognition for the
(24:51):
next generation. The science isclear. The brain needs effort,
friction, handwriting, andreflection to grow strong. If we
hand over all the heavy liftingto AI, we might find that we've
traded ease for erosion. A bitof housekeeping.
We were nominated for a Women inPodcasting Award for the second
year in a row. We're so excitedabout it. Thank you so much. The
(25:14):
link to vote for the show willbe in the show notes, and I
would love your vote for BestScience Podcast. As always,
until next time, stay curious,and don't forget to subscribe so
you never miss a week.
Popular Podcasts
United States of Kennedy
United States of Kennedy is a podcast about our cultural fascination with the Kennedy dynasty. Every week, hosts Lyra Smith and George Civeris go into one aspect of the Kennedy story.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.