Post
JA EN

Is 'Hardships in Youth Are Worth Buying' Outdated in the AI Era? — The Surprising Answer from Neuroscience and Learning Science

Posted
By
15 min read
Is 'Hardships in Youth Are Worth Buying' Outdated in the AI Era? — The Surprising Answer from Neuroscience and Learning Science
  • Target audience: IT engineers interested in career and learning strategies for the AI era
  • Prerequisites: None
  • Reading time: 15 minutes

Overview

“Hardships in youth are worth buying” — this is a well-known Japanese proverb (wakai uchi no kurou wa katte demo seyo) that encourages young people to actively seek out challenges and difficulties. But in an age where AI is taking over intellectual tasks, is this just outdated motivational rhetoric? The evidence from neuroscience and learning science is surprisingly consistent: appropriate cognitive load is an essential investment for the brain. The critical shift, however, is that the quality of struggle, not the quantity , has fundamentally changed. This article examines research on cognitive reserve, desirable difficulties, and neuroplasticity to scientifically distinguish the struggles worth pursuing from those worth avoiding in the AI era.

The Brain Physically Grows Through Struggle

The Brains of London Taxi Drivers

Some of the most compelling evidence that cognitive load physically reshapes the brain comes from a famous study of London taxi drivers.

Maguire et al. at UCL (2000) 1 found that London taxi drivers, who memorize the city’s intricate road network, had significantly larger posterior hippocampi — the brain region responsible for memory and spatial cognition — compared to the general population. Crucially, there was a positive correlation between years of experience as a taxi driver and hippocampal volume.

These results alone could not rule out the possibility that people with naturally larger hippocampi were simply more likely to become taxi drivers. To address this, Woollett & Maguire (2011) 2 conducted a longitudinal study tracking taxi driver trainees over four years. The results were unambiguous — every trainee who passed the notoriously difficult exam known as “The Knowledge” showed increased gray matter in the posterior hippocampus, while those who failed and the control group showed no change. Cognitive challenge physically altered brain structure.

Juggling and Gray Matter

A similar phenomenon has been observed in juggling. Draganski et al. (2004) 3 reported in Nature that practicing juggling increased gray matter in brain regions involved in visual-motor processing. However, these changes reversed when practice stopped — supporting the “use it or lose it” principle.

flowchart TB
    A["Engage in cognitively<br>demanding activities"] --> B["Neural circuits are<br>strengthened (synaptic plasticity)"]
    B --> C["Brain structure<br>physically changes"]
    C --> D["Cognitive reserve<br>accumulates"]
    D --> E["Greater resilience against<br>age-related cognitive decline"]

    F["Offload cognitive<br>effort to external tools"] --> G["Neural circuit<br>usage decreases"]
    G --> H["Brain structure<br>may shrink"]
    H --> I["Cognitive reserve<br>does not accumulate"]
    I --> J["Greater vulnerability to<br>age-related decline"]

    style A fill:#4CAF50,color:#fff
    style E fill:#2196F3,color:#fff
    style F fill:#f44336,color:#fff
    style J fill:#FF9800,color:#fff

Cognitive Reserve — How Early Investment Protects Your Aging Brain

What Is Cognitive Reserve?

The “Cognitive Reserve” theory proposed by Yaakov Stern at Columbia University 4 explains why individuals with similar levels of brain pathology can have vastly different outcomes in terms of dementia onset. Cognitive reserve, accumulated through intellectual activity, acts as a buffer against brain damage.

According to Stern’s 2012 paper in The Lancet Neurology 5, individuals with fewer than 8 years of education had a relative risk of developing dementia that was 2.2 times higher than those with more education.

Overwhelming Evidence from Meta-Analyses

Valenzuela & Sachdev (2006) 6 conducted a systematic review of over 29,000 participants and found that high cognitive reserve — measured by education, occupational complexity, and participation in intellectual activities — was associated with a 46% reduction in dementia risk (odds ratio 0.54, 95% CI 0.49–0.59).

Furthermore, the Lancet Commission report (2024 update) 7 identifies 14 modifiable risk factors that account for approximately 45% of all dementia cases, listing low educational attainment as a major risk factor in the early stages of life.

In other words, challenging your brain while young is literally an insurance policy that can halve your future dementia risk.

Desirable Difficulties — Not All Struggle Is Created Equal

Bjork’s Desirable Difficulties Framework

Here is where an important distinction must be made. Not all struggle is equally beneficial.

The concept of “Desirable Difficulties,” originally proposed by UCLA’s Robert Bjork in 1994 and reviewed in 2020 8, demonstrates that conditions that feel difficult during learning actually enhance long-term retention and transfer substantially.

Three classic examples of desirable difficulties are:

1. The Testing Effect (Retrieval Practice)

Roediger & Karpicke (2006) 9 compared students who took practice tests with those who simply re-read the material. On a short-term test 5 minutes later, the re-reading group performed better. But after 2 days and 1 week, the testing group significantly outperformed them. In other words, the struggle of retrieving information from memory is precisely what strengthens it.

2. Spaced Practice (The Spacing Effect)

Cepeda et al. (2006) 10, in a meta-analysis covering 317 experiments and 839 assessments, confirmed that spaced practice is overwhelmingly more effective for long-term retention than massed practice. The “inconvenient” approach of spacing out repetitions over time works far better than cramming.

3. Interleaving

Rohrer & Taylor (2007) 11 showed that mixing different types of problems during practice produces better learning outcomes than practicing one type at a time. The confusion may feel counterproductive, but it promotes deeper cognitive processing.

Undesirable Difficulties

On the other hand, the following are “undesirable difficulties” that hinder learning:

  • Meaningless repetitive tasks (e.g., manual copy-and-paste work)
  • Learning under excessive stress (fear and intense pressure impair memory through cortisol)
  • Tasks far beyond the learner’s current level (attempting advanced work without foundational knowledge)
  • Spending time on simple tasks that can be automated

This distinction is the crux of the struggle debate in the AI era.

What “Struggles Worth Buying” Look Like in the AI Era

Just as GPS Eroded Spatial Memory, AI Risks Eroding Thinking Ability

Dahmani & Bohbot (2020) 12 found that habitual GPS users had lower spatial memory, and a 3-year longitudinal follow-up (a small sample of n=13) showed that hippocampus-dependent spatial memory declined more rapidly. Sparrow et al. (2011) 13 reported the “Google Effect” in Science — when people expect information to be available online, their memory encoding for that information decreases.

Just as GPS altered spatial cognition and Google changed how we encode memories, AI tools pose a broader risk of cognitive offloading.

The Cost of “AI Makes It Easy” — A Large-Scale PNAS Field Experiment

Bastani et al. (2025) 14 reported a large-scale field experiment in PNAS that scientifically validates this concern. In a study of approximately 1,000 high school students learning mathematics:

  • Students with access to GPT-4 improved their practice performance by 48%
  • However, when AI access was removed, they performed 17% worse than students who never used AI at all
  • When educational guardrails were in place (providing hints rather than answers), this negative effect was substantially mitigated

This finding is critically important. Using AI as a “tool that gives you answers” reverses the learning effect. However, using AI as a tool that supports the thinking process does not impair learning.

Young Brains Are Especially Vulnerable

The prefrontal cortex — responsible for planning, impulse control, working memory, and attention — continues to mature until around age 25 15. If cognitive challenges are insufficient during this developmental period, it could affect the formation of neural circuits that are still under construction.

Gerlich (2025) 16, in a study of 666 participants, found that frequent use of AI tools was negatively correlated with critical thinking ability. The 17–25 age group showed the highest levels of AI dependency and consistently lower critical thinking scores compared to other age groups.

So What Should Young People “Buy” in the AI Era?

Synthesizing the research evidence, we can organize it as follows.

Struggles Worth Pursuing (Scientifically Beneficial Cognitive Load)

Type of StruggleScientific BasisPractical AI-Era Examples
Thinking through problems independentlyDesirable difficulties 8, retrieval practice 9Form your own hypothesis before consulting AI. Spend at least 30 minutes working independently before asking AI for help
Tackling diverse problemsInterleaving effect 11Don’t stay confined to one technology — alternate between problems from different domains
Spaced repetitionSpacing effect 10Use spaced repetition systems (SRS) instead of cramming
Metacognitive reflectionCognitive reserve accumulation 4Ask yourself: “Why did this approach work?” and “What alternatives exist?”
Teaching and explaining to othersGeneration effect and elaboration 17Restate AI-generated answers in your own words
Venturing into unfamiliar territoryNeuroplasticity 12Step outside your comfort zone and explore new paradigms

Struggles to Avoid (Unproductive Load Best Left to AI)

Type of StruggleReasonAI Use Case
Writing boilerplate codeRepetitive work that doesn’t promote cognitive growthDelegate code generation and focus on design decisions
Manually debugging syntax errorsMechanical tasks with minimal learning valueLet AI identify errors; invest time in understanding root causes
Exhaustive information retrievalSearch itself contributes little to cognitive growthUse AI to aggregate information; focus on critical evaluation
Formatting and administrative tasksRequires no higher-order cognitive skillsAutomate these and redirect time to creative and analytical work

Growth Mindset and Its Relationship to Struggle

Mueller & Dweck (1998) 18 demonstrated across six experiments that children praised for effort sought out challenges, while children praised for intelligence avoided them. A mindset that frames struggle not as “evidence of inadequacy” but as “part of the growth process” plays a crucial role in cognitive development.

Moser et al. (2011) 19 used EEG to show that individuals with a growth mindset exhibited heightened neural attention to errors (increased error positivity/Pe), which mediated improvements in post-error accuracy. In other words, people with a growth mindset literally learn more from mistakes at the neural level.

Applied to the AI era, this means that rather than passively accepting AI’s “correct” answers, the process of personally verifying them and discovering errors is what trains the brain.

Deliberate Practice for Expertise Remains Unchanged

Ericsson et al. (1993) 20 showed that individual differences in expert performance are more closely associated with accumulated deliberate practice than innate talent. Deliberate practice is defined as effortful activities performed at the edge of one’s current abilities, specifically designed to optimize improvement.

This principle holds true in the AI era. What changes is the content of deliberate practice.

Traditional StruggleDeliberate Practice in the AI Era
Writing algorithms from scratchVerifying and improving AI-generated code
Reading documentation cover to coverGetting AI summaries, then diving deep into critical design decisions
Writing unit tests manuallyFocusing on test strategy design and edge case discovery
Googling error messagesDeveloping a systemic understanding of root causes

Conclusion — Struggle Hasn’t Become Unnecessary; It Has Evolved

“Hardships in youth are worth buying” — it turns out this proverb is not outdated motivational rhetoric after all.

Neuroscience is clear: cognitive challenges physically reshape brain structure 123, build cognitive reserve 456, and reduce future dementia risk by 46% 6. Learning science has repeatedly demonstrated that “struggling while learning” substantially improves long-term retention (the testing effect has been shown to roughly double retention rates) 8910. That said, the magnitude of these effects varies between individuals, and the same results are not guaranteed for everyone.

However, the old proverb needs an update.

The proverb for the AI era: “Appropriate cognitive load in youth is worth buying — but what counts as ‘appropriate’ has changed.”

In an age where AI handles simple cognitive tasks, there is no value in suffering through menial work. But if we abandon the struggle of thinking just because AI can provide answers, we risk losing our most essential cognitive abilities 14 — much like how habitual GPS users lose their spatial memory 12.

This is especially critical for those under 25, whose prefrontal cortex is still developing 15. It is precisely during this period that challenging the brain through appropriate cognitive demands represents the greatest investment in a lifetime of cognitive health.

There is one more important point. Cognitive investment doesn’t end in youth. Hultsch et al. (1999) 21 showed in a 6-year longitudinal study that people who continued intellectual activities in middle and later life experienced slower cognitive decline. The “principal” of cognitive reserve built in youth needs continued intellectual challenge to protect its “interest” over time. While the specific types of cognitive investment change with age, the principle of maintaining appropriate cognitive load persists throughout life.

AI should not be used as a “tool that eliminates struggle” but rather as a “tool that elevates the quality of struggle”. Delegate simple tasks to AI and redirect the freed-up time toward higher-order cognitive challenges — design decisions, critical thinking, creative problem-solving. That is what science supports as the “struggle worth buying” in the AI era — not just for young people, but for everyone.

For more on related topics, check out these articles:

References

References corresponding to the citation numbers in the text are listed below in numerical order.

Additional References (not cited by number in the text)

  1. Navigation-related structural change in the hippocampi of taxi drivers - Maguire, E.A. et al., Proceedings of the National Academy of Sciences, 97(8), 4398-4403 (2000). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  2. Acquiring ‘the Knowledge’ of London’s layout drives structural brain changes - Woollett, K. & Maguire, E.A., Current Biology, 21(24), 2109-2114 (2011). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  3. Neuroplasticity: Changes in grey matter induced by training - Draganski, B. et al., Nature, 427(6972), 311-312 (2004). 【Reliability: High】 ↩︎ ↩︎2

  4. What is cognitive reserve? Theory and research application of the reserve concept - Stern, Y., Journal of the International Neuropsychological Society, 8(3), 448-460 (2002). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  5. Cognitive reserve in ageing and Alzheimer’s disease - Stern, Y., The Lancet Neurology, 11(11), 1006-1012 (2012). 【Reliability: High】 ↩︎ ↩︎2

  6. Brain reserve and dementia: a systematic review - Valenzuela, M.J. & Sachdev, P., Psychological Medicine, 36(4), 441-454 (2006). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  7. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission - Livingston, G. et al., The Lancet, 404(10452), 572-628 (2024). 【Reliability: High】 ↩︎

  8. Desirable difficulties in theory and practice - Bjork, E.L. & Bjork, R.A., Journal of Applied Research in Memory and Cognition, 9(4), 475-479 (2020). Original: Bjork, R.A. (1994). Memory and metamemory considerations in the training of human beings. In Metacognition: Knowing about knowing, MIT Press. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  9. Test-enhanced learning: Taking memory tests improves long-term retention - Roediger, H.L. & Karpicke, J.D., Psychological Science, 17(3), 249-255 (2006). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  10. Distributed practice in verbal recall tasks: A review and quantitative synthesis - Cepeda, N.J. et al., Psychological Bulletin, 132(3), 354-380 (2006). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  11. The shuffling of mathematics problems improves learning - Rohrer, D. & Taylor, K., Instructional Science, 35, 481-498 (2007). 【Reliability: High】 ↩︎ ↩︎2

  12. Habitual use of GPS negatively impacts spatial memory during self-guided navigation - Dahmani, L. & Bohbot, V.D., Scientific Reports, 10, 6310 (2020). 【Reliability: Medium-High】 ↩︎ ↩︎2

  13. Google effects on memory: Cognitive consequences of having information at our fingertips - Sparrow, B. et al., Science, 333(6043), 776-778 (2011). 【Reliability: High】 ↩︎

  14. Generative AI without guardrails can harm learning: Evidence from high school mathematics - Bastani, H. et al., Proceedings of the National Academy of Sciences, 122(26) (2025). 【Reliability: High】 (Note: An erratum has been published. DOI: 10.1073/pnas.2518204122) ↩︎ ↩︎2

  15. Maturation of the adolescent brain - Arain, M. et al., Neuropsychiatric Disease and Treatment, 9, 449-461 (2013). 【Reliability: High】 ↩︎ ↩︎2

  16. AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking - Gerlich, M., Societies, 15(1), Article 6 (2025). 【Reliability: Medium】 (Note: A correction was published in September 2025; there are methodological concerns. Causal inference is limited due to the cross-sectional design) ↩︎

  17. The generation effect: A meta-analytic review - Bertsch, S. et al., Memory & Cognition, 35(2), 201-210 (2007). 【Reliability: High】 ↩︎

  18. Praise for intelligence can undermine children’s motivation and performance - Mueller, C.M. & Dweck, C.S., Journal of Personality and Social Psychology, 75(1), 33-52 (1998). 【Reliability: Medium-High】 ↩︎

  19. Mind your errors: Evidence for a neural mechanism linking growth mind-set to adaptive posterror adjustments - Moser, J.S. et al., Psychological Science, 22(12), 1484-1489 (2011). 【Reliability: Medium-High】 ↩︎

  20. The role of deliberate practice in the acquisition of expert performance - Ericsson, K.A. et al., Psychological Review, 100(3), 363-406 (1993). 【Reliability: High】 ↩︎

  21. Use it or lose it: Engaged lifestyle as a buffer of cognitive decline in aging? - Hultsch, D.F. et al., Psychology and Aging, 14(2), 245-263 (1999). 【Reliability: High】 ↩︎

Career
Cognitive-Science neuroscience AI Learning Career Cognitive-Reserve neuroplasticity
This post is licensed under CC BY 4.0 by the author.