Academic Integrity & AI

Academic Integrity in UG Mathematics at Warwick

A vital aspect of your studies now and in the future is academic integrity, and this document aims to give you an overview of what that means both in general and for your studies in Mathematics at Warwick in particular. In a nutshell, academic integrity means avoiding all forms of dishonesty and cheating in your work, demonstrating intellectual ownership of your work, and acknowledging the sources that your work builds upon. It is important that you take the time to understand these issues, because breaches of academic integrity can have serious negative consequences for your studies.

We understand that these issues may be new to you, and that the mention of University regulations may be scary. You may well have questions about academic integrity. If so, then please do ask your personal tutor: they will be happy to help, and a discussion about these issues among your tutor group, in which you think through what it all means in a few example situations, can help to clarify things.

The information in this guide is intended to be fairly informal and readable; for more formal information, including the University's regulations on the topic, see Regulation 11 Academic Integrity.

You are also encouraged to complete the following Moodle courses:

What is Academic Integrity?

Academic integrity means a fundamental commitment to honesty in all aspects of one's academic work. It is perhaps easiest to define academic integrity by what it is not. There are a few main behaviours that breach or fall short of the expected standards of academic integrity.

Cheating

An obvious way to breach academic integrity is to cheat in an assignment, test, examination. The term cheating covers obvious offences such as copying in tests or exams, stealing work from other students (either electronically or in another way), or taking unauthorised materials (notes/phones/smart watches etc.) into an exam venue. It is also cheating to assist another student to cheat (e.g. by allowing them to copy your work) — indeed, this can be considered collusion. Cheating is dealt with as a serious offence.

Plagiarism

You may know the saying, attributed to Isaac Newton, "If I have seen further, then it is by standing on the shoulders of giants." Newton is referring to the fact that as scientists we inevitably use the prior works of other people in our own works. However, doing so without giving credit where it is due is plagiarism — presenting another's work as your own. The "other" here could be a fellow student, a textbook, or a generative artificial intelligence — the principle is the same in each case.

Superficial modification of source material does not make it your own work, so one can still plagiarise even without it being a case of exact one-for-one copying. The acknowledgement of sources using appropriate citations is the difference between plagiarism (bad) and scholarship (good).

The same is true of joint work, where the proper acknowledgement of co-authors and their contributions is essential. In Mathematics at Warwick, we encourage discussion of assignment problems amongst students, but the submitted work should be fundamentally your own. If in doubt, simply acknowledge discussions before writing out your version (e.g. "In discussions with A. N. Other we found the following method for solving this problem...").

"Avoiding Plagiarism" Moodle Course

Falsification

By falsification we mean the presentation of statements that are outright untrue, in a dishonest or misleading way.

A simple slip of the pen, accidentally converting $e^{i \pi} + 1 = 0$ (true) into $e^{i \pi}-1=0$ (false), is not falsification nor is it a breach of academic integrity.
However, fabricating a data set for use in an essay or paper, and then claiming that it resulted from a genuine scientific experiment, would be a case of falsification.
Conversely, generating a synthetic data set for the purposes of testing an algorithm, and reporting that the algorithm passed this test (with a clear statement that it was synthetic data), is normal and good practice.
Similarly to plagiarism, the dishonest modification of source material or data can also be a form of falsification.

Because falsification undermines the very notion of truth in science, and hence society's confidence in scientific processes and institutions, it is taken very seriously.

AI and AI: Academic Integrity and Artificial Intelligence

In some sense, the advent of generative artificial intelligence changes everything: here is a technology that can do research, and impressive feats of analysis and reasoning, at the push of a button — and at significant environmental cost! It is certainly tempting to use it in our learning and research.

However, the expectations of academic integrity remain the same: sources must be cited, help and co-authorship must be acknowledged (and may be judged to be excessive even when acknowledged), and it is essential that the human author submitting work for credit has intellectual ownership of the work.

In this sense, generative AI is just another resource, perhaps more technological than a book in the library, but not fundamentally different. If you do use AI in your work, then follow the same good practices as you would with any other source:

Clearly set out why, where, and how you have done so. This may be done with a clear paragraph in your submission. Specific citation requirements may be given to you as part of any assessment brief and, if so, then follow those instructions.
Keep a clear record of all interactions with any AI that you later use or rely upon for a submission. This may be via screen-grabs, PDF print-outs, or other recording techniques. Again, some assignments may have specific instructions for you to follow.
The Library has a Moodle course on artificial intelligence and academic integrity, which you may wish to consider completing.

One should also bear in mind that generative AI can fabricate information and findings (so-called "hallucinations") — in academic integrity terms, it can commit falsification.

Where is the line?

One can naturally ask where the line really lies between making use of humanity's stores of knowledge on the one hand, and breaching academic integrity on the other hand. There are two significant factors to bear in mind:

Any attempt to deceive or dishonestly gain an advantage is a breach of academic integrity. That said, one can also breach academic integrity without deliberate intent (although that lack of intent could be a mitigating factor).
In work that builds upon other works — and, as Newton pointed out, this is often unavoidable — the central issue is not whether the work is solely your own, but whether you have intellectual ownership of the work. Simply put, are you the author of the work, the synthesiser of the acknowledged ideas of others, or are you merely a copy-paster of items from elsewhere?

So, how does one avoid even unintentional deception, demonstrate intellectual ownership, and practice academic integrity in mathematics?

Academic Integrity in Mathematics and at Warwick

Mathematical Culture

The expectations of academic integrity regarding cheating, plagiarism, and falsification are more-or-less universal. However, the expectations of academic integrity in the mathematical sciences can differ somewhat from those of our colleagues (and your friends) in other departments, particularly when it comes to citations.

Where our communities agree is that in research papers and books, all sources must be cited — to do otherwise would be to implicitly and falsely claim the copied material as one's own, which would be plagiarism. A social scientist might achieve this by an in-line citation such as "see Smith (1988, p.207)" and the full bibliographic details of Smith's 1988 paper or book will appear as a footnote or endnote. Mathematicians would be more likely to refer to a result than a page, e.g. as "see Smith (1988, Theorem 12.4)" or use a numerical referencing style, e.g. "see [15, Theorem 12.4]".

Citations should be clear in both directions:

It must be clear which part of your work is covered by the citation, so place citations appropriately, e.g. at the end of the relevant sentence or paragraph.
It must be clear which part of the cited work is being cited. It is not enough to simply write "see Smith (1988)" when that is a 1000-page book and one really only means Theorem 12.4. That said, if the work being cited is particularly short, or one is genuinely referring to the work as a whole, then citations like "the definitive work in this area is the book of Smith (1988)" may be appropriate — but, generally, one should try to be specific.

Where our communities diverge is on the citation of what might be called "classical" material. A lot of what you will learn in undergraduate mathematics is literally centuries old. Your Analysis lecturer is almost certain to call the sum rule for limits of sequences simply "the sum rule" with no reference to who first proved it or when. Even for famous named results, we mathematicians generally find it enough to write e.g. "Taylor's theorem" — no-one except a historian of mathematics would cite the exact work of Brook Taylor's in which it first appeared, and in any case the modern form of the result differs from the original. Our colleagues in other fields might cite classical material in the same way as recent papers, but we generally do not, and especially not in lectures (although textbooks might cite more).

(By the way, the sum rule appears to have been first proved with rigour by Augustin-Louis Cauchy in his 1821 Cours d'Analyse.)

Academic Integrity in Mathematics at Warwick

What are the expectations of academic integrity in Mathematics at Warwick? There is something of a spectrum.

At the most rigorous end, we expect written examinations to be free of any attempt to cheat, and breaches of this will be treated most seriously.
At the least rigorous end, we understand that solutions to weekly assignment sheets on classical areas of mathematics are not going to be extensively cross-referenced, and this expectation is the usual one in the mathematics community. We also understand that these problem sets will probably be discussed in supervision and tutor groups — actually, we encourage this. Nevertheless, the solutions that you submit for credit should still fundamentally be your own work, and if this is not the case, then you should at least declare it. Also, at the end of the day, if you don't do the work yourself during term-time, then you are not actually mastering the mathematical skills — and the truth of what you understand and what you don't will shine through in your exam performance. As George Pólya put it, "mathematics is not a spectator sport".
In between these two extremes lie lengthier written pieces of work such as your first year group project, second year essay, and fourth year project. Here, we expect to see proper citation of sources and the contributions of co-authors, with some allowances made for learning how to do this in your first year.

To repeat what was mentioned earlier: If you have any questions about this, then ask your personal tutor, who will be happy to help and facilitate a discussion around these topics. You can also take the "Avoiding Plagiarism" Moodle course.

Breaches of Academic Integrity

Processes

Now we turn to the unpleasant business of breaches of academic integrity. Because academic integrity underpins the value of our teaching, learning, and research, breaches of academic integrity are serious matters. Although they are not disciplinary matters at the University of Warwick, they are still handled through formal processes that can affect marks and ultimately degree outcomes. These processes are spelled out in detail in Regulation 11, and this section gives an overview of how things work in Mathematics.

Minor breaches of academic integrity are called poor academic practice. These are typically dealt with by the marker of the affected piece of work and result in a reduction of the mark, often arrived at by setting aside the affected portion of the work and awarding the mark that the unaffected portion deserves. On the other hand, more serious breaches are called academic misconduct, and they attract more serious penalties. In general, work tainted by academic misconduct receives a mark of zero.

If concerns are raised about a piece of your work, then those concerns will be considered independently by the Department's Academic Integrity Team. If they decide that there is enough evidence, then a meeting of an Academic Conduct Panel (ACP) will be called. Two or more members of the Academic Integrity Team will consider all the relevant evidence, including anything that you wish to present. The ACP will make a decision on the balance of probabilities, and a recommendation to the Head of Department:

It may be decided that the evidence does not prove the case, and that there is, after all, no case to answer and no penalty to apply.
It may be decided that academic integrity standards have been breached, but in a relatively minor way that constitutes poor academic practice. This outcome often results in a lower mark that is based on the quality of the unaffected portion of the work.
It may be decided that academic integrity standards have been breached in a serious way that constitutes academic misconduct. This outcome often leads to a mark of zero for the affected work.
Severe, aggravated, or repeated cases may be escalated to the University's Academic Integrity Committee, which has the power to impose greater sanctions than the Department does. These include revoking the right to remedy failure, or recommending withdrawal from the University.

Please be aware that, for the purpose of the Regulations, the Mathematics Department deems that the wording "the piece of work concerned" in the Regulations means:

the whole of the assessed component of a module for which the majority of the credit comes from an examination (so cheating on a single homework assignment or class test may mean that the department sets to zero the marks on the whole assessed component of a module);
all the work submitted for a single deadline for those modules which are more completely assessed (so that, for example, cheating on any part of an assessment may mean that the department sets to zero the marks on the whole of that assessment, even if it is broken into several parts.)

Help and Advice

If you do ever find yourself involved in an academic integrity process, then you have the right to seek advice and support from your personal tutor, the Students' Union, and indeed anyone else. One such person may accompany you to the ACP meeting to help and support you, but they cannot speak on your behalf. If you disagree with the ACP's outcome, then you may appeal to the University's Academic Integrity Committee.