The Old Model and Why It’s Under Pressure

Essays and homework were long the stalwarts of assessment. They measure knowledge, writing skills, and critical thinking. But with the presence of AI, it is now easy to produce well-written essays, finish problem sets, or even codes in minutes.

That does not mean students are learning less—it’s just that the tools they use have changed. Relying on the old model without adapting is like asking students to write out multiplication tables manually once calculators are employed everywhere. It’s not getting it.

 Redesigning Exams

Exams are designed to test individual knowledge. When AI is introduced, we may need to:

• Shift from recall to reasoning: Instead of “What happened in 1857?” ask “How might the outcome of the 1857 revolt have changed if modern communication technology existed?” This tests creativity and analysis, not memorization.
• Use open-book / open-AI exams: Allow students to use tools but focus on how well they apply, critique, and cross-check AI’s output. This mirrors real-life work environments where AI is available.
• In-person oral or viva testing: Requiring students to orally discuss their answers tells you whether they actually understand, even if they had AI help.
• Timed, real-world problem-solving: For math, science, or business, create scenarios that require quick, reasonable thinking—not just memorization of formulas.

Testing is less “what do you know” and more “how you think.”

 Rethinking Projects & Coursework

Projects are where AI may either replace effort or spark new creativity. To keep them current:

• Process over product: Teachers need to grade the process—research notes, drafts, reflection, even the mistakes—not just the polished final product. AI can’t get away with that iterative process so easily.
• AI within the assignment: Instead of banning it, design assignments that require students to show how they’ve used AI. For example: “Employ ChatGPT to generate three possible outlines for your paper. Compare them, and explain what you retained and what you eliminated.”
• Collaborative assignments: Group work encourages skills AI finds it difficult to replicate well—negotiation, delegation, creativity in group work.
• Hands-on or practical elements: A project assignment could be an interview of grandparents, a science project would be the making of a small prototype. AI must complement but not replace lived experiences.

This reverses coursework from being outsourcing-oriented to reflection, uniqueness, and human effort.

 Reframing Coursework Purposes Altogether

If AI is already capable of doing the “garden variety” work, maybe education can focus on more higher-order goals :

• Critical thinking with AI: Are students able to recognize errors, biases, or gaps in AI-generated work? That’s a skill used in the real world today.
• Authenticity and voice: AI can generate text, but it can’t replicate the lived experience, feeling, or creative individuality of a student. Assignments could emphasize personal connections or insights.
• Interdisciplinary study: Promote projects that combine math, art, history, or ethics. AI is good at doing one thing, but human learning thrives at points of convergence.

The Human Side

This’s not about “catching cheaters.” It’s about recognizing that tools evolve, but learning doesn’t. Students want to be challenged, but also supported. When it all turns into a test of whether they can outsmart AI bans, motivation falters. When, on the other hand, they see AI as just one of several tools, and the question is how creatively, critically, and personally they employ it, then education comes alive again.

 Last Thought

Just as calculators revolutionized math tests, so will AI revolutionize written work. Prohibiting homework or essays is not the answer, but rather reimagining them.

The future of exams, project work, and coursework must:

• Distrust memorization more than thinking, applying, and creating.
• Welcome AI openly but insist on reflection and explanation.
• Strive for process and individuality as much as product.
• Retain the human touch—feelings, experiences, collaboration—at its center.

In short: assessments shouldn’t try to compete with AI—they should measure what only humans can uniquely do.