Student Cited for Misconduct for Using Grammarly

Plus, more fines and stupid questions about yet another penalty for cheating accountants. Plus, from Monitor AI.

Issue 263

To join the 3,725 smart people who subscribe to “The Cheat Sheet,” enter your e-mail address below. New Issues every Tuesday and Thursday. It’s free:

If you enjoy “The Cheat Sheet,” please consider joining the 15 amazing people who are chipping in a few bucks via Patreon. Or joining the 28 outstanding citizens who are now paid subscribers. Thank you!

---

Student Cited for Misconduct for Using Grammarly

According to an outlet called Distractify, a student has been cited for misconduct and been given a zero on her college assignment for using Grammarly. Though not reported directly in the coverage, the student attends the University of North Georgia.

Before I dive in, let me say how much I loathe — yes, loathe — the idea of writing about a story that’s based entirely on what someone posted on TikTok. I’m doing it anyway because I think the issue is important, and having risked the brain cell loss to actually watch the student’s performance outrage, it’s decently sourced. There are names of people who appear to actually exist.

So, here we are.

The skinny version is that a junior at UNG turned in a paper for a criminal justice course and received a zero. The assessment was issued due to her use of Grammarly and its AI generative technology. The student says she did not know Grammarly used AI and that its use has been allowed and even encouraged by other faculty. The professor, it seems, has not relented and initiated a misconduct case against the student. It’s also noteworthy that it appears that Grammarly has joined the issue, commenting on the student’s page that they were contacting the school regarding its policies.

And here we go.

One, Grammarly is a generative AI tool that suggests and allows people to use AI-generated text in place of their own writing. Simple. Fact. True.

Many people, including me, have raised this issue (see Issue 249). In it, I wrote:

I suspect many educators have not absorbed that tools such as Grammarly, Google Translate and Quillbot are AI-powered text changers that change student writing. That is to say that they write for students. They, and ChatGPT, are different fruit in the same tree.

So, does using Grammarly allow AI to stand-in for authentic work? Yes. No question. As such, if a professor does not want students to use AI to re-do their writing, they are well within bounds to penalize that conduct.

Professors should be clear about these expectations early on, including mentioning products such as Grammarly and Quillbot. They should make it clear that using them is a form of cheating and therefore disallowed. In this particular case, we don’t know whether those expectations were set.

Though here I think it’s important to note that the students says she spoke with the teacher as well as the “department head and the dean” about the issue — which is what we want. Conversation. Dialogue. And, hopefully, learning.

In the same video, the student says, regarding Grammarly:

even though it wrote absolutely nothing for me, I still get the zero.

This cannot be true. Clearly, the student used Grammarly. She says she did. Grammarly changed something in her paper. Absent that, there would be no way to tell. And again, when Grammarly or any of these “checkers” replace your words for theirs, they are writing for you. This does not feel like some arcane concept beyond the comprehension of a college student.

She says repeatedly as the foundation of her anger that she did not know Grammarly used AI to do its work. Maybe that’s fair. But I have to ask, where did she think Grammarly came up with its suggestions? I mean, it’s clearly not a human tutor making editing suggestions, right? It’s very obviously computer created text. More on point, it does not matter. Once you take text from someone or something and include it in your work without citation, representing it as your own, that’s plagiarism. Not maybe. That’s what that word means. The AI is ancillary.

So, if you used Grammarly or Quillbot to swap its words for yours, it’s exceptionally likely that someone — maybe many someones — will consider that to be cheating. Accordingly, there may be consequences.

Our student, clearly not getting this, incredulously tells the universe at one point:

It’s Grammarly!

The implication being that Grammarly should be OK because, you know, it’s Grammarly. It only tells you what words and punctuation to use instead of your own. No problem.

To me, the pervasive normalization of these checkers and automated tutors and paraphrasers is a real problem. I have no doubt that Grammarly can write a clean paper. The question was - can you?

The reason Grammarly is involved here is pretty obvious. Our student started her multi-episode TikTok performances by advising people to uninstall Grammarly so they don’t get nailed for cheating like she did:

If you have a paper, and it’s getting turned in to Turnitin, uninstall Grammarly right now

Grammarly does not want people to uninstall their product. To work, Grammarly use has to be normal and routine and not cheating. No wonder they’d want to try to convince the school that using generative AI to modify a student’s work is not cheating. I can hear their slogan now — it’s not cheating, it’s Grammarly! Overlay magic wand sound and pixie dust graphics.

The student has her misconduct hearing in early January. I will try to remember to follow up.

Another Fine for Major Accounting Firm for Exam Cheating — Examinees Thought Open Sharing of Answers was “Collaborative Culture”

The news is that a regulator in Ontario has fined major accounting and auditing firm PwC just over $1 million for widespread exam cheating.

There’s a link to the citation in this article from Growing Concern, which I use as the source here because it has a good, albeit brief, rundown of the absolute mess that accounting firms have created for themselves with exam cheating. From the article:

Ever since KPMG was fined $50 million by the SEC for cheating in 2019 it’s been a free-for-all on the answer sharing punishments around the world. EY US got fined $100 million by the SEC for doing it in 2022 though theirs was a bit worse as they were also cheating on the ethics exams one must take to be licensed as a CPA and not just WBLs.

And just a few weeks ago, the PCAOB announced it was fining PwC China and PwC Hong Kong a total of $7 million USD for their own cheating: From 2018 until 2020, over 1,000 individuals from PwC Hong Kong, and hundreds of individuals from PwC China, engaged in improper answer sharing – by either providing or receiving access to answers through two unauthorized software applications – in connection with online tests for mandatory internal training courses related to the firms’ U.S. auditing curriculum.

And that’s not even all of it.

On this infraction at PwC specifically, the reports show that:

From at least 2016 to early 2020, more than 1,200 PwC Canada personnel were involved in improper answer sharing related to training tests.

Twelve hundred. In Canada.

From the regulator in Ontario:

There was a consistent mindset among the participants that engaged in Answer-Sharing that it was both widely-known and appropriate. Many viewed sharing answers as part of a collaborative culture at PWC and because the assessments were open book, some did not view answer sharing as ethically improper.

I say this again — unless you actively do something to deter or mitigate cheating, people in your orbit will come to believe you accept and promote it. Having a policy on the books is not enough. Turning the blind eye is permission.

Before moving on, I will also share this story-ending blurb from Growing Concern:

We feel compelled to ask — does anyone other than regulators give a sh*t if some overworked auditors share answers on open book, checkbox assessments? Truly?

I edited the expletive.

I care. We all should. And I do not understand people who don’t.

It matters that auditors and accounts may not actually know the material and may not be able to actually perform the tasks their employer is selling. It matters if you’re buying those services. It matters that companies and professions — auditors especially — have a culture of corner cutting. It matters if accounts feel it’s appropriate to focus exclusively on the outcome and ignore the rules and established processes. Competency and integrity matter. Heaven help us if they don’t.

From AI Monitor

I am always happy to open this space to allies — the people and companies engaged in protecting the integrity, quality and fairness of education. If you have news or ideas to share, please let me know. A reply e-mail works.

The below is from our friends at AI Monitor, unedited:

Maffei Consulting Group, LLC (MCG) believes educating future generations is one of the most noble goals that one can ever endeavor upon, for education is the march of progress, lifting up the next generation to stand upon our shoulders so that they may see further and dream bigger than we could have ever imagined. With the rise of AI, we quickly saw challenges to education the likes of which have never been faced.

It was out of these new AI imposed challenges to education and the need for a reliable, ethical, and equitable solution to those challenges that our small team began working diligently to bring educators a responsible solution for the detection of AI plagiarism, culminating in the creation of AI.Monitor.

AI.Monitor works by providing a comprehensive analysis of a student’s writing process thereby identifying deviations from typical academic writing patterns, this novel approach, gives educators clear and understandable insights into whether or not the work submitted by their students was the student’s own or was the creation of AI or another. AI.Monitor’s patent pending approach to  the detection of AI plagiarism provides educators with a responsible solution, the benefits of which are:

1.      Future-Proof Design – AI.Monitor is designed knowing that AI is dynamic and ever evolving. Entirely new AI models, like ChatGPT 4, are already being released while at the same time existing AI models are autonomously improving their performance. With each new model and update the resulting AI generated output is becoming more human like, this is why AI.Monitor is focused on detecting AI plagiarism through what remains human, the process of writing itself. This novel method for detecting AI plagiarism provides educators with reliable functionality independent of new AI model or autonomous updates, making AI.Monitor a solution which works today, tomorrow, and far into the future.

2.      Transparent & Understandable Method – With AI.Monitor, there are no black-box algorithms, it’s straightforward, a student either put forth the effort required to create their paper, or they did not. This simplicity provides educators with a transparent and understandable solution to detecting AI plagiarism, as all the metrics being examined and analyzed are easily accessible to them.

3.      Omni-Lingual Capability – Creating an ethical and equitable AI plagiarism detector requires a solution which is functional in every language and for people of all linguistic backgrounds. AI.Monitor is designed with these cultural sensitivities at the forefront, allowing educators to have a solution that does not perpetuate historical inequitable access to technology, ensuring that linguistically diverse individuals are not marginalized through biased AI plagiarism detection. It does not matter if you say “Guten Tag!” for hello or “Adios!” for goodbye because AI.Monitor provides the same level of functionality for writers of any language ensuring the same experience regardless of language used.

4.      Privacy Focused – We are staunch supporters of privacy at MCG and believe that the least amount of data necessary to perform a task should be the only data used. That’s why AI.Monitor does not collect or store any student’s writings, allowing for full functionality while maintaining total privacy. Some things should stay out of the hands of giant corporations, a student’s work is one.

5.      Protected Detection Methodology – AI.Monitor is equipped with two views, a Student View and an Instructor View. The Student View collects metrics on a student’s writing process in real time while concealing those metrics from them. The Instructor View provides educators an easy to use assessment of a student’s writing pattern as well as detailed documentation of the metrics used for AI.Monitor’s analysis. This two view design hinders the ability for determined students to bypass AI.Monitor’s AI plagiarism detection capability while providing educators full insight into their student’s writing process.

6.      Extremely Affordable – AI.Monitor is a subscription which costs only $29.99 a year per student. The cost can be covered entirely by students, split between the institution and the students, or completely covered by the institution, providing flexibility to the individual circumstances of the institution and the students.

If you are considering AI.Monitor for your classroom or have any questions at all, I encourage you to email me at joshmaffei@aimonitor-mcg.com or visit our website at strategymcg.com/aimonitor. We at MCG are happy to help you in any way we can whether that is a discussion on best practices or implementing AI.Monitor in your classroom!

Class Note

A reminder that the much-anticipated, though somehow also under-appreciated, Best and Worst of 2023 is coming soon. Please send nominations, ideas or thoughts by reply e-mail. For review, here are the Best/Worst Issues from 2022 and 2021.