AI vs AI in Education: AI Should Extend Thinking, Not Replace It

A growing concern in education is not simply that students are using AI. It is how they are using it.

When a student asks a model to write the essay, create the analysis, formulate the argument, and polish the conclusion, the student may submit competent-looking work without doing the intellectual work that education is supposed to develop.

Writing is not only a way to display knowledge. It is one of the ways we form it.

The uncomfortable process of choosing words, organizing evidence, recognizing weak arguments, revising assumptions, and defending a conclusion is where much of the learning happens. Remove all of that friction, and AI can become less of a learning tool and more of a cognitive outsourcing system.

That is the issue behind the current wave of concern around AI-assisted assignments. But I do not think the answer is simply to ban AI and pretend it will disappear. I think we need to distinguish between AI that replaces thinking and AI that creates conditions for better thinking.

The difference matters

There is a major difference between these two prompts:

"Write my essay about the fall of Rome."

and:

"Take me to Rome in 410 CE. Let me speak with people experiencing the collapse of imperial authority. Show me competing perspectives, force me to make decisions with incomplete information, then challenge my assumptions."

The first produces an answer. The second creates a learning experience. That distinction is central to how I think about learning with Airtrek.ai.

Airtrek is not being built as a machine that hands someone the final interpretation of history. The goal is to use AI, narrative, simulation, and interactive environments to place learners inside the complexity that textbooks often flatten.

History is rarely just a list of dates, leaders, wars, and outcomes. It is uncertainty, competing accounts, limited information, cultural context, power, survival, tradeoffs, and human consequence. An immersive AI experience can make that complexity more accessible, but it must not make the learner passive.

How I approach learning in Airtrek

My approach is built around active inquiry rather than automated answers. Instead of asking an AI system to tell a learner what to think, Airtrek can prompt the learner to investigate:

• What does this person know, and what can they not know?
• Whose perspective is missing from the record?
• What evidence supports one interpretation over another?
• What changes if you look at the event through the eyes of a farmer, trader, soldier, child, leader, or displaced family?
• What assumptions are you bringing from the present into the past?
• What would you do with only the information available at that moment?

The AI becomes an interactive guide, historical role-player, research companion, and source of productive friction, not a replacement author. A learner might enter a historically grounded environment, speak with several simulated perspectives, encounter conflicting evidence, form a hypothesis, and then be challenged to explain why they believe it.

That final step matters. Airtrek should not reward someone for accepting the first coherent-sounding answer. It should reward the ability to ask better questions, compare claims, identify uncertainty, and revise an interpretation when the evidence changes. That is a very different learning loop from "generate, copy, submit."

What China's approach signals

China's school AI strategy is interesting because, at least in its public policy framing, it does not treat AI literacy as merely teaching children how to prompt a chatbot. Its curriculum direction includes AI concepts, data and algorithmic thinking, practical skills, ethics, critical evaluation, and awareness of AI's limits. It also recognizes risks such as dependency, plagiarism, and uncritical trust in generated output. That is closer to the direction education needs.

The future should not divide students into two groups: those who use AI and those who do not. AI will be too embedded in work, information systems, and everyday life for that to be realistic. The essential divide will be between people who can use AI without surrendering their judgment, and people who cannot.

Not built for kids, yet

Airtrek is not currently designed as a children's education product. The immediate focus is on building an intelligent, immersive learning environment that can support exploration of history, culture, place, and human experience. But I do see a future route toward age-appropriate educational experiences.

Any future version intended for younger learners would need to be designed around clear guardrails:

• AI should ask questions before providing conclusions.
• Experiences should encourage exploration, observation, and discussion.
• Claims should be traceable to sources, with uncertainty made visible.
• Learners should be prompted to explain their reasoning in their own words.
• Adult educators and parents should retain visibility and control.
• The system should avoid turning historical simulation into false certainty or persuasive role-play detached from evidence.

The goal would not be to create young "prompt engineers." It would be to help develop historically aware, curious, skeptical, empathetic, and capable thinkers who understand both the power and the limitations of AI.

The educational opportunity

There is a real danger in treating AI as a shortcut around difficult thinking. But there is also a real opportunity to use it to make difficult thinking more compelling.

A learner may not want to read a disconnected paragraph about an ancient city, a migration, an archaeological site, or a political collapse. But they may become deeply engaged when they can explore the place, ask questions, encounter ambiguity, compare perspectives, and discover that history was lived by real people facing uncertain choices.

That engagement is not a substitute for literacy, writing, research, or analysis. It is an invitation into them.

The best educational AI will not say, "Here is the answer." It will say:

"What do you notice? What evidence supports that? What might you be missing? What changed your mind?"

That is the path I see for Airtrek: not automating education, but making the work of learning more immersive, meaningful, and human. Read more about how this works at How Airtrek Creates Immersive Historical Experiences and History, Interpretation, and AI in Airtrek.

References

• Landymore, Frank. "Students Across the Country Are Using AI Models to Complete Assignments and Write Entire Essays." Futurism, August 2026.
• Goldstein, Dana. "What Happens When Students Let A.I. Write Their Papers?" The New York Times, August 17, 2026.
• Kosmyna, Nataliya, et al. "Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task." MIT Media Lab research, 2025.
• Australian Government, Department of Education. "China's Approach to AI Education in Primary and Secondary Schools." September 2025.