A Practical Guide to Education in an AI World

Ikhlaq Sidhu

AI is here - and it’s here to stay.  We've only begun to see what's possible. AI is an incremental innovation for every creator and a disruptive force to every incumbent. In the area of education, AI has introduced new questions, problems, and solutions. We have been told that AI will improve the teaching-to-student ratio and that the quality of education will also improve.  However, many practical questions have surfaced, including:

1. Should students be allowed to use AI to do their homework?
2. If we don't allow AI in education, isn't it a disservice to students who won't learn to use modern tools and thus be uncompetitive in the job market?
3. Is it cheating to use AI in a group project?
4. Should faculty use AI to create teaching content?
5. Is it cheating to use AI on a test?
6. What if students use AI for homework and teachers use AI for grading—what’s the point?

As educators, we swing back and forth on solutions. One day, we say, “let students use the most modern tools”. The next day, we say, “let's cut off their internet during exams”. In sum, we need a framework for Education and AI – and the framework described below can help educators find that balanced approach.

Axioms for AI in Education:

Let's start with a few axioms. You may know that there are 10 axioms that form the basis of math today. In math, an example axiom is 1=1, it can not be absolutely proven, but that common sense tells us are unlikely to be disputed. There were originally 9 axioms, and then later they realized they need to add a 10 one to make the rest of math work out. In case you are interested, the 10th axiom says that that between every 2 numbers, there exists another number. All of our advanced math today is built upon these 10.

Here are my axioms of AI for Education:

1. AI is a tool and an assistant for everyone.
2. Students can and should use AI to help them whenever they can, especially in learning, but not necessarily in testing.
3. Faculty can and should use AI whenever they can, especially in functions that let them scale-up their work or improve its quality.
4. The learning process has two major classifications, with a grey area between them: Fundamentals and Applications.
5. Learning Fundamentals includes learning to read, write, comprehend, and do basic math. More advanced topics involve theories and laws, like F = ma in physics. Fundamentals can be learned with or without AI. 
6. Application of Fundamentals is when we combine multiple fundamental areas to produce a work product.
7. Testing of Fundamentals should be without AI, because some things just need to be memorized or internalized. In the same vein, grade school students wouldn't be allowed to use calculators to answer 4+5.
8. Testing of Applications should allow AI. This is when we test the ability of a student to apply fundamental knowledge, possibly from different subjects. Application is often worked on in teams and advanced tools are often allowed. Whether writing a marketing plan or a PHD thesis, AI should be allowed. However, projects can and should be larger, more complex, or more open-ended when AI tools are expected to be used.
9. The grey area will be an ongoing debate about what constitutes Fundamentals vs. Applications. From practice, we know that as topics become complex, there are often two tests: one for the fundamentals without tools and another that might be open-book, open-calculator, and now also open-AI.

Note that even in the High School IB programs (International Baccalaureate) of today, there are 2 math tests. One which is completed on paper and pencil without a calculator and the other may be taken with a calculator. This idea of 2 or more assessments for one subject is not new!

Why These Axioms?

We aim to enable students to succeed in the next generation and sharpen the skills they’ll need in a world where AI acts as an assistant and intermediary. When calculators were introduced, teachers thought it would be the end of math education. However, over time, we’ve learned that a valuable set of skills is needed to:

1. Understand what problem is worth solving.
2. Frame the problem for the computation.
3. Interpret the result.
4. Use critical thinking to test the solution.
5. Take actions because of the computation.

Moving forward, AI is replacing the calculator - not just with calculations but also with computation that includes writing, summarizing, generating, simulating, and more. However, the human parts—ranging from the entrepreneurial intuition of recognizing the problem to critical thinking about the answers and then taking the next step—are actually more important than executing the computation.

In Conclusion

We need students to:

1. Know the fundamentals so they can use and recall them on demand.
2. Be able to apply advanced tools while augmenting their complementary and inherently human capabilities.

These 9 axioms provide a starting framework for how to use AI to enhance education. Maybe later, we will also need a 10th one.