Using AI (Gemini) to Study Better — A Guide for Returning Students

What this is

A practical guide to using an AI assistant (this one is written around Google Gemini, but the method works with any good AI) to learn, not to cheat. It is written especially for someone going back to study after years away — for example, a first-year Computer Science student at Unisa taking Introduction to Computer Science, Calculus, and Algebra.

It gives you a small set of copy-paste prompts you can reuse every day, plus a study routine that works for distance learning.

The golden rule: Use AI as a tutor, not an answer machine. If you only copy answers, you will feel productive and then freeze in the exam. If you use it to understand, you will pass and actually know the work.

Why it matters

Going back to study after 10 years out of school is hard, but it is also a superpower — adults are more disciplined and more motivated than 18-year-olds. The two things that usually trip up returning students are:

1. Maths confidence — it feels rusty, so calculus and algebra look scarier than they are.
2. Distance learning loneliness — at Unisa, nobody stands in front of you. You set the pace. There is no one to ask "what does this mean?" at 9pm.

A good AI fixes exactly these two problems. It is a patient tutor that never judges you for asking a "stupid" question, available 24/7, that can explain the same thing five different ways until it clicks. Used correctly, it is the closest thing a distance student has ever had to a private teacher.

The knowledge

Step 1 — Turn on the right mode

Gemini has a built-in study mode called Guided Learning (launched August 2025, powered by Google's LearnLM models, which are tuned specifically for education). Instead of giving you the final answer, it walks you through a problem step by step, asks you questions, and checks if you understood. This is the mode you want for studying.

• In the Gemini app, look for Guided Learning (under tools / "Gem" options).
• Gemini can also make study guides, flashcards, and quizzes from your own notes or textbook chapters. You can build a quiz in Canvas and choose how many questions and whether they are multiple-choice or short-answer.
• A Gemini Ultra / Pro subscription gives full access to these tools and the strongest models — use them.

If Guided Learning is ever not available, you can force the same behaviour with a prompt (see the "tutor prompt" below). The mode just saves you typing it every time.

Step 2 — The four reusable "tutor prompts"

Save these four. They are the whole system. Paste, fill in the blank, go.

1. The Tutor prompt (use this to start any topic)

You are my patient tutor. I am a first-year university student who has been
out of school for 10 years, so explain in simple language and don't assume
I remember school maths. Topic: [PASTE TOPIC OR QUESTION].

Don't give me the full answer yet. First explain the idea in plain words with
a simple real-life example. Then give me ONE small practice question and wait
for my answer before continuing.

2. The "I'm stuck" prompt (use on a hard problem)

Here is a problem I'm stuck on: [PASTE PROBLEM].
Don't solve it for me. Give me only the FIRST hint or first step,
then ask me what I think comes next. Go one step at a time.

3. The "check my work" prompt (use after you try yourself)

I tried to solve this myself. Here is my attempt: [PASTE YOUR WORK].
Check it. If it's wrong, tell me WHERE I went wrong and WHY, but don't
rewrite the whole thing — let me fix it myself. If it's right, give me a
slightly harder version to test if I really understand.

4. The "explain it again, differently" prompt (when it doesn't click)

I still don't understand [CONCEPT]. Explain it a completely different way —
use a picture-in-words analogy, then a step-by-step example with real numbers.
Then ask me a question to check I got it.

Step 3 — Prompts for each of your modules

Introduction to Computer Science / Programming (e.g. Unisa COS1511, Introduction to Programming I)

The biggest mistake in a first programming course is asking AI to write the code for you. You learn programming by struggling and fixing, like learning to drive. Use AI to explain, not to type.

Explain what this code does line by line, in plain English, like I've never
programmed before:
[PASTE CODE]

My code has an error. Here is the code and the error message. Don't fix it
for me — explain in simple words what the error message MEANS and what part
of the code to look at:
[PASTE CODE AND ERROR]

Give me a beginner practice exercise about [TOPIC, e.g. loops / if-statements
/ arrays]. Just the exercise, not the solution. After I send my attempt,
review it.

Tip: If your course uses a specific language (Unisa COS1511 uses C++), tell the AI the exact language so its examples match what you submit.

Calculus (Unisa MAT1512 — Calculus A, a year module)

I'm starting calculus and my school maths is rusty. Before we do calculus,
test my basics: give me 3 quick questions on [fractions / algebra basics /
functions] and tell me what to revise first based on my answers.

Explain what a [derivative / limit / integral] actually IS — what it means in
real life — before any formulas. Use a simple example like speed and distance.
Then show one worked example slowly.

Here is a calculus problem. Give me one hint at a time, don't solve it all:
[PASTE PROBLEM]

Algebra (Unisa MAT1503 — Linear Algebra I, a year module — expect vectors, matrices, and systems of equations)

Teach me [topic, e.g. solving simultaneous equations / matrices / factorising]
from the very beginning, assuming I forgot it. Plain words first, then one
worked example, then give me one to try.

Show me, step by step, how to check my own algebra answer so I don't need to
rely on you. What can I substitute back in to know if I'm right?

Step 4 — A weekly routine that works for distance learning

Distance study fails when there's no rhythm. Build a small fixed one:

1. Read first, ask second. Read the Unisa study guide / textbook section before opening Gemini. Then use AI to clear up what you didn't get. AI fills gaps; it doesn't replace reading.
2. One topic per sitting. Pick one idea (e.g. "for-loops" or "the chain rule"), use the Tutor prompt, do 2–3 practice questions. Stop. Small wins build confidence.
3. Always try before you peek. Attempt the problem on paper first, then use the "check my work" prompt. The struggle is where learning happens.
4. End each week with a quiz. Ask Gemini to make a 10-question quiz on the week's topics. Mark yourself honestly. Your wrong answers are next week's study list.
5. Past papers are gold. When you have past exam papers, work through them yourself, then use the "check my work" prompt. This is the highest-value thing you can do before exams.

Things to watch out for

1. AI gets maths wrong sometimes. This is the big one. AI can confidently give a wrong calculation or a wrong final number. Always verify maths yourself — substitute the answer back, or re-check the steps. Treat AI as a study partner who is usually right but not perfect, not as a calculator. (For pure arithmetic, ask it to "show every step" — it's more reliable that way — or use an actual calculator.)

2. Never submit AI-written work. Unisa (like all universities) treats AI-generated assignments as plagiarism / academic dishonesty, and it can get you disqualified. Use AI to understand, then write the assignment and the code in your own words and your own hands. If you can re-explain it without AI, you've learned it. If you can't, you've only copied.

3. Don't let AI replace the official material. Your marks come from the Unisa study guide, tutorial letters, and prescribed textbook. AI explains those better — it doesn't replace them. The exam tests their content, not the AI's.

4. Don't ask for the answer too early. The moment you ask "just give me the answer," you've stopped learning. Use the "one hint at a time" prompts.

5. Protect your confidence. If a topic feels impossible, that's normal after 10 years away — it's not a sign you can't do it. Tell the AI "explain it like I'm starting from zero" and go slower. Everyone's brain remembers maths once it warms up.

6. Keep your real questions for your real tutors too. Unisa has e-tutors, discussion forums, and lecturers. AI is the 24/7 backup; the official support is still yours to use.

Where to learn more

On Gemini's study features:

• Google blog — How Gemini's Guided Learning can help you study
• Google blog — New Gemini tools for students (Aug 2025)
• Google Workspace Updates — New study tools in the Gemini app

On Unisa Computer Science modules:

• Unisa — Computer Science subjects & modules
• Likely first-year modules: COS1511 (Introduction to Programming I, in C++), MAT1512 (Calculus A), MAT1503 (Linear Algebra I). Confirm the exact codes on myUnisa → registered modules, and read each module's Tutorial Letter 101 first — it tells you everything about assignments and the exam.
• Module pages: Linear Algebra – MAT1503

Family knowledge holders:

• Anyone in the family who has studied programming or maths — ask them. A real person plus AI beats either one alone.