Using AI to learn, not to skip the learning: a guide for business students

Every student I teach now has a generative AI tool in their pocket. The question is no longer whether they will use it, but whether it will make them sharper or lazier. I hold NPTEL certification in Artificial Intelligence in Human Resource Management, and I teach students to use AI critically: as an aid to thinking, never a replacement for it. This guide sets out the habits I think make the difference.

The real risk is not cheating

Copying an AI’s answer into an assignment is the risk everyone talks about, and it matters. But a quieter risk matters more. Learning happens in the struggle: in outlining an answer badly and then better, in getting a problem wrong and finding out why. An AI that removes the struggle removes the learning, even if no rule is broken.

A business graduate will be paid for judgment: deciding what a set of numbers means, what a customer really wants, which of three plans to back. Judgment is built by practice. Outsource the practice and the judgment never arrives.

Four habits that keep you the author

The figure above shows the loop. Each step has a reason.

1. Attempt it first. Before you open the AI, write your own outline, attempt the calculation, or note what you already know. Even a weak attempt gives you something to compare against, and shows you exactly where you are stuck.

2. Ask a sharp question. Give context and ask for reasons. Not “explain break-even analysis”, but “I’m a second-year B.B.A. student. Here is my break-even calculation; I got 1,200 units but the answer is 1,000. Where is my mistake? Explain the step, don’t just give the answer.” Other useful requests:

• “Ask me five questions on Herzberg’s theory, one at a time, and tell me what I missed.”
• “Give me a different example of unity of command being broken, from a service business.”
• “Here is my answer. What would an examiner say is missing?”
• “Argue against my conclusion as strongly as you can.”

3. Check every claim. AI tools produce fluent text, and fluency is not accuracy. They can get numbers wrong, mix up theories, and invent sources: articles, authors and court cases that look real and do not exist. Check every fact against your textbook, the original source or a reliable website. Never cite a reference you have not found and read yourself.

4. Rewrite in your own words, and say where AI helped. Close the AI and write the final answer yourself. If you cannot, you have not learned it yet; go back to step 2. Where your teacher allows AI, a one-line note (“I used an AI tool to check my calculation and to generate practice questions”) is honest and increasingly expected.

What not to give an AI

• Personal data: your own or anyone else’s phone numbers, addresses, ID numbers or health details.
• Confidential material: data from an internship or a company project, unless the organisation has approved the tool.
• Your thinking, before you have done it. The order matters: you first, then the AI.

Prompting is a business skill too

Writing a good prompt is much like writing a good brief for a colleague: state who you are and what you need, give the context, specify the format, set the limits, and ask for reasons. Students who learn to do this well, and to check what comes back, are learning a skill employers increasingly look for. That is a better use of AI in a business degree than any shortcut.

For teachers

The habits above also suggest how to set work in an AI era: ask for the draft as well as the final answer, ask students to critique an AI’s answer rather than produce one, use viva-style questions on submitted work, and design tasks rooted in local, specific contexts (a business in your own town, data the class collected) that a generic model cannot simply answer.

None of this needs a ban, and none of it needs blind enthusiasm. It needs the same thing good teaching has always needed: keeping the student doing the thinking.