Why Teach AI With a Fixed Curriculum?
AI changes every few weeks. Education should adapt to the learner at the same speed.
A school recently invited me to enrol in a 13-month graduate program in artificial intelligence.
My first thought was simple:
Which version of AI will they be teaching in month thirteen?
AI changes too quickly for a fixed course. Models improve. Tools disappear. New capabilities replace techniques people spent months learning.
The school is trying to stay relevant. The problem is not the course material.
The problem is the course.
A course is built before the student arrives. The school decides what should be taught, in which order, at what pace, and for how long.
The student then fits the curriculum.
AI gives us the opportunity to reverse this model.
AI responds to context
Ask AI what it does and there is no single answer.
It writes software for a developer.
It explains a financial statement to a business owner.
It tutors a student.
It reviews a contract for a lawyer.
It helps an architect assess a design.
The underlying system might be the same. The outcome changes because the context changes.
Context includes the person’s objective, experience, history, documents, instructions, constraints, and current problem.
AI does not begin with a fixed workflow.
It begins with:
What are you trying to accomplish?
This is why AI is disruptive.
Traditional software requires the person to adapt to the product.
AI adapts the product to the person.
STOP! Read that again!
AI adapts the product to the person.
Education should do the same.
Curriculum is designed for the average student
A curriculum must serve many students at once.
It defines a common starting point, sequence, pace, assessment, and finish line.
One student already understands software architecture.
Another has never written code.
Another understands machine learning but lacks business experience.
They still enter the same program.
The experienced student repeats material.
The new student struggles to keep pace.
The curious student finds an important question outside the syllabus and is told to return to the assigned lesson.
The curriculum manages the student.
It does not understand the student.
Context should replace the fixed path
Education still needs outcomes.
An AI engineering student should understand model behaviour, data, evaluation, architecture, security, governance, and production delivery.
But every student does not need to reach those outcomes through the same path.
The system should begin by understanding the learner’s context:
• What do they already know?
• What are they trying to achieve?
• Which knowledge gaps block their progress?
• Which examples fit their experience?
• How quickly are they learning?
• What has changed in the field since yesterday?
An experienced cloud architect should not spend weeks learning basic cloud concepts.
A marketing executive should not receive the same technical path as a software developer.
A student building an AI system for construction should learn through construction plans, building codes, document analysis, and regulatory risk.
The destination might be shared.
The path should belong to the student.
The project becomes the curriculum
Curiosity follows personal relevance.
You encounter a problem.
You ask a question.
The answer reveals another gap.
You follow it because the knowledge matters now.
Suppose you are building an AI system to review construction plans.
You soon need to understand document extraction, image analysis, retrieval, privacy, model reliability, human review, audit records, cloud costs, and legal accountability.
The project determines the sequence.
Your existing knowledge determines the pace.
Your questions determine the depth.
This is a curriculum assembled from context.
It changes as the learner changes.
It also changes as AI changes.
The role of the school must change
AI does not remove the need for education.
It changes what the institution should provide.
The school should define credible outcomes.
It should assess the student’s current capabilities.
It should provide expert judgment, reliable sources, guardrails, feedback, practical experience, and trusted assessment.
AI should provide the adaptive learning path.
The teacher is no longer responsible for repeating the same lesson to every student.
The teacher challenges the student’s assumptions, evaluates their work, identifies blind spots, and confirms whether they have developed the required capability.
The institution stops selling access to information.
It starts validating understanding.
A course is a product
This is the same mistake organizations make everywhere.
An airline focuses on the seat instead of the person who must sit in it.
A bank focuses on the account instead of the person trying to manage their money.
A school focuses on the course instead of the person trying to learn.
The product exists first.
The customer arrives later.
AI reverses this relationship because its behaviour emerges from context.
That gives education a choice.
Schools can place AI inside the existing course and call the program modern.
Or they can use AI to rebuild education around the student.
The first option protects the course.
The second serves the learner.
Why sell a fixed course about a technology that changes every few weeks?
Because the institution knows how to build courses.
That does not mean the course is the best way to learn.
The future of education is not a better curriculum.
It is a learning system that understands the person well enough to know what they need next.