News

Universities set rules on where AI is allowed in studying

Instead of bans, more universities ask for one thing: be honest about where a neural network helped.

For you: Before your term paper, check your university's rules and note in the paper where AI helped.

Universities set rules on where AI is allowed in studying

A year ago many lecturers simply banned neural networks and ran assignments through detectors. The detectors were often wrong, and students used chat assistants anyway. The ban did not work: it pushed AI help into the shadows and cast suspicion on students who wrote everything themselves.

Now universities, including several in Kazakhstan, are moving to clear rules. The question is no longer “did the student use AI” but “how did they use it, and were they honest about it”. Here is what is usually allowed, what is not, and how to avoid losing a pass over a couple of lines you forgot to write.

Why the bans failed

A detector for generated text cannot read minds. It looks for statistical signs: sentences that are too smooth, word choices that are too predictable. But real people write like that too, especially those writing in a second language or following a template carefully. As a result, the detectors were often wrong: they flagged honest students and missed those who lightly reworded a chatbot's answer.

The second reason is even simpler. A chat assistant sits on every phone, and students used it no matter what. A rule that cannot be checked stops being a rule. So lecturers shifted from hunting for cheaters to setting an agreement: help is fine here, and cheating starts there.

What is usually allowed

The list differs from one university to another, but the typical set looks like this:

  • explaining a difficult topic — breaking down a formula, a concept or a theory in plain words;
  • making a plan — for a term paper, a presentation or an essay;
  • finding mistakes in a text — grammar, typos, gaps in logic;
  • practising before an exam — asking the assistant to quiz you on a topic and check your answers.

The logic is simple: help with explanations is treated like a session with a tutor. If you asked a senior student or a lecturer, nobody would call it cheating. The same goes for a neural network, as long as you do the thinking and the writing.

Here is a request that usually fits the rules:

Explain [topic] in simple words, as if to a first-year student. Give two real-life examples. Then ask me three questions to check that I understood.

What is usually not allowed

The bans are also similar across most universities:

  1. Submitting generated text as your own. This is the main rule. Handing in someone else's text counts as cheating, whether a classmate or a model wrote it.
  2. Using AI in an exam without permission. If the lecturer has not said clearly that you may, assume you may not.
  3. Citing sources that do not exist. Neural networks can confidently invent books, articles and authors. If such a reference ends up in your paper, you are responsible for it, not the model.

Be especially careful with the last point. Check every reference a chatbot gives you: does the book exist in a library or catalogue, and does it really contain that page and that quote. If you cannot verify a reference, leave it out.

The key requirement: transparency

What is new in these rules is not the list of bans but the demand for honesty. At the end of a paper, students add a couple of lines on which tool they used and why. For example:

In preparing this paper I used [tool name] for [understanding the topic / making a plan / checking grammar]. The text, conclusions and choice of sources are my own.

Lecturers say that an honest note rarely lowers the grade. Hidden help, once discovered, can cost you a pass. The maths is easy: the note costs nothing, while silence can cost a lot.

There is another benefit. When you write down where AI helped, you learn to separate your own work from the tool's work. That skill stays useful after university: at work you will also need to explain what you did and what the software did.

What to do now if you study in Kazakhstan

The rules are still taking shape, and they can differ between universities and even between lecturers. So act ahead of time:

  1. Find your university's rules. Look on the university website, in the academic integrity policy, in the course syllabus. If there is nothing, ask the lecturer directly, ideally in writing.
  2. Check the rules for each subject. One lecturer may allow a plan made by a neural network, another may allow only grammar checks.
  3. Keep your drafts and chat history. If questions come up, you can show how the work developed.
  4. Verify every fact and source from a model's answer against real books and articles.
  5. Add a note at the end of your term paper, essay or report, even if the help was small.
  6. Use AI to understand, not to hand in. If after a chat with the bot you can explain the topic in your own words, you have learned something. If not, the model did the work for you.

Neural networks are not leaving education. The real choice is whether you use them openly and to your own benefit, or risk a pass to save a few hours.

Ad Your ad hereA banner on the homepage, in stories and sections. Formats and prices on our Advertise page.

Comments

No comments yet. Be the first.