Design Thinking in the Age of AI: Why Schools Need It Now

By HCD Institute

There is a quiet, measurable cost to how most people now solve problems: they open a chat window and ask. A 2025 MIT Media Lab study fitted essay writers with EEG caps and found that the more a writer leaned on an LLM, the weaker their brain's own neural connectivity became — the group that wrote unassisted showed the strongest, widest-ranging networks; the group that used AI showed the weakest. A separate peer-reviewed study of 666 participants, published in Societies (2025), found a statistically significant link between frequent AI tool use and declining critical-thinking scores, driven by what researchers call cognitive offloading — the habit of handing a mental task to a machine so consistently that the underlying skill quietly atrophies.

None of this means AI is bad. It means that thinking, like a muscle, gets weaker when it stops being used — and the tools that make life easiest are, by default, the ones most likely to be overused. The question worth asking is not should we use AI but what discipline keeps a human mind sharp while using it. Design thinking is one of the few methodologies built, from first principles, to answer that question.

AI Is Fast at Answers. It Cannot Frame the Question.

Design thinking's first move is not to solve — it is to empathise: to sit with a real person's problem before touching a solution. That single habit is precisely what generative AI erodes, because AI is built to shortcut it. Ask a chatbot for ideas for a new school orientation programme and it will hand you ten plausible-sounding ideas in four seconds — none of them grounded in a single conversation with an actual anxious new student. The ideas feel like progress. They are actually the illusion of progress, because the hardest and most valuable part of problem-solving — deciding which problem is actually worth solving — has been skipped.

This is the trap researchers are now documenting directly: teams get seduced by AI's speed, rushing to solutions before truly understanding the problem. Design thinking's discipline — empathise, define, ideate, prototype, test, in that order, with real users at each step — is a structural defence against exactly this trap. It does not reject AI. It insists that a human still owns the framing, the fieldwork, and the final judgment, while AI accelerates the parts in between: pattern-finding in research data, generating a wider spread of prototype directions, simulating edge cases before a human tests the real one.

The Two Things AI Cannot Do For You

Two capabilities sit at the centre of design thinking, and both are, not coincidentally, the two capabilities every serious study of AI's cognitive impact flags as most at risk.

Original thinking. An LLM's output is, definitionally, a recombination of what has already been written. It is extraordinarily good at producing the average good idea instantly. It is structurally unable to produce the strange, specific, first-hand observation a designer makes by actually watching a farmer struggle with a form, or a child give up on a worksheet. Design thinking trains people to generate that observation themselves — which is also, functionally, training them to keep exercising the exact cognitive muscle the MIT and Societies studies found weakening under AI dependence.

Empathy. A chatbot can produce fluent, even moving, empathetic-sounding language. It has no stake in the person on the other end and no access to what they didn't say. Design thinking's empathise phase is not a writing exercise — it is fieldwork: interviews, observation, sitting in someone else's discomfort long enough to see the problem the way they see it. That is a human skill, and like any skill, it needs practice starting young, not a late-career workshop after twenty years of default AI use has already set in.

Why This Has to Start in Schools, Not the Boardroom

Most organisations that adopt design thinking do it with adults who already have twenty years of habitual problem-solving behind them — good and bad. The habits are harder to change than to build correctly the first time. That is the real argument for starting in schools, and it is now backed by a growing body of K-12 research: studies using empathy maps in classroom design-thinking projects found students across a wide age range able to build genuine empathetic understanding of a problem, and a broader review of design-thinking-integrated learning found consistent gains in creativity, collaboration, and — specifically — critical thinking, developed together rather than as separate subjects.

India's own policy architecture already points this direction. The National Education Policy 2020 explicitly calls for a shift away from rote memorisation toward experiential, inquiry-based learning, with design thinking named as a mechanism for building exactly the competencies — creativity, critical thinking, collaboration, communication — that cognitive-offloading research shows are most vulnerable to erosion in an AI-saturated environment. A student who has spent years running real empathise-define-ideate-prototype-test cycles on problems that matter to them enters an AI-filled world with the habit of framing problems and forming original judgments already built in. A student who has not spent those years is starting that training from zero, at the exact moment AI makes it easiest to never start at all.

HCD Labs: A Design Lab Inside the School, Not a One-Off Workshop

This is the thinking behind HCD Labs, the schools programme HCD Institute runs under the Design Innovation Centre framework at IIT Hyderabad. Rather than a single workshop, HCD Labs establishes a permanent, university-grade design studio inside a partner school — for clay modelling, materials exploration, electronics prototyping, and AI-assisted design work — for students aged 12 to 18, with annual IIT-faculty-led teacher certification and a direct pathway for exceptional students into IIT programmes. It is built to be NEP 2020-aligned by design, and it treats AI as one tool on the bench among many, not the starting point for every assignment.

The premise is simple: if original thought and empathy are the two things AI cannot replace, they are also the two things worth deliberately practising before AI use becomes the default. A design lab inside a school, running for years rather than a single term, is where that practice compounds.

The Takeaway

AI is not going to make problem-solving obsolete. It is making the undisciplined version of problem-solving — jump straight to an answer, skip the fieldwork, accept the first plausible output — dramatically easier to fall into, at precisely the moment research shows that pattern measurably weakens critical thinking. Design thinking is not a defence against AI. It is the discipline that makes AI genuinely useful, by keeping a human firmly in charge of the two things no model can do: deciding which problem is real, and understanding, first-hand, the person who has it. The earlier that discipline is built — ideally in school, not after twenty years of habit — the more it holds up.

Frequently Asked Questions

Does design thinking work against using AI tools? No — it changes the order of operations. Empathise and define happen with humans first; AI is used inside the ideate and prototype stages to widen options and speed iteration, not to replace the framing.

What age should design thinking training start? K-12 research shows students across a wide age range can meaningfully build empathy and critical-thinking skills through design projects; HCD Labs is built for ages 12–18, aligned to secondary-stage NEP 2020 guidance.

Is the evidence on AI and critical-thinking decline solid? The MIT Media Lab study is a preprint and not yet peer-reviewed, so its conclusions are preliminary; the Societies (2025) cognitive-offloading study is peer-reviewed and points the same direction. Both warrant caution rather than alarm — and both are consistent with why deliberate, human-led problem-solving practice matters.

Key Takeaways

• Peer-reviewed and preprint research both link heavy AI reliance to weaker critical thinking, through a mechanism called cognitive offloading.

• Design thinking's empathise-define-ideate-prototype-test discipline structurally protects the two things AI cannot replicate: original, first-hand observation and genuine empathy.

• Starting the practice in school, before AI-first habits set in, is more effective than retraining adults after twenty years of default use.

• NEP 2020 already points India's schools toward experiential, design-led learning — the same competencies the AI research shows are most at risk.

• HCD Labs builds this as a permanent school studio under IIT Hyderabad's Design Innovation Centre, for ages 12–18.

Mathew Sebastian

Mathew is a mentor at HCD Institute, where he has led the movement to democratise design thinking in India since 2011. An alumnus of NID Ahmedabad and a Fellow at IIT Hyderabad's Design Innovation Centre, he brings over 18 years of experience across design strategy, education, and public policy.
He has advised governments and institutions including the Government of Kerala, Bihar's Ministry of Industries, the Andaman & Nicobar Administration, and Nordic diplomatic missions — with a curriculum formally adopted by Mahatma Gandhi University.