Vol 3, Issue 14. Quarter 3 – 2026. 

A Professor at Alcorn State University became notable recently due to a bit of pedagogical sleight of hand. He gave his students a question in an electronic document. What he did not tell them was that he had embedded a word in the document that the students couldn’t see because it used a font that was the same color as the background. However, when the text was copied into a chatbot, the AI engine could  “see” the word “Madagascar” in the middle of the question. When ChatGPT generated answers to the question it almost universally included the word Madagascar in the response, even though it had nothing to do with the question being asked. 

32 out of his 35 students submitted their responses that included phrases such as, “Madagascar floats sideways through the afternoon,” and “Madagascar  purple bicycle whispers to the ceiling.” Clearly any student who proofread their work before handing it in would have caught such nonsense, but apparently virtually none of them did. Many onlookers online congratulated the professor for teaching the students valuable lessons in honesty.

But here is the thing. Why was AI not allowed to begin with? As a professor, I was one of those folks who came up with all kinds of tricks to prevent chatbots from being used on tests and similar assignments. The easiest way was to simply restrict the test to paper and a pencil while sitting in my presence. I also made multiple versions of the exams so that the person next to you had different data than you did, and the problem seemed to be solved. But consider this. 

Consider a random student, who I will call Joe. Joe is approaching his college graduation. He applies for a job by responding to an ad generated by AI. Joe drafts a cute cover letter and submits a resume.  The company gets thousands of these resumes every week, so it uses AI to generate the profile of a perfect applicant for each job. The AI engine scores each resume based on how close it comes to this ideal. HR uses these scores to decide which resumes to show to the manager looking to fill the position. It then uses AI to generate a common bank of questions for the manager to use in interviews to ensure that the company is being “fair” to all of the applicants. A transcript is generated from the interview and AI scores the responses. This generates hiring recommendations. Joe is the lucky student that gets the job and shows up for work. On day 1 he sits in some meeting looking at slides generated by AI. Later he reviews the meeting by looking at the summary generated by AI. After that he works through his to-do list generated by AI, and he is expected to get these tasks done instantly by using AI to help him create an efficient “Gem” which integrates an optimized workflow. After a few months of this Joe gets a letter, generated by AI telling him what a great job he is doing.

Now we have to ask the big question. Which student is better off at this point: Fred, the student who did the reading in the English class, or Joe who learned how to use AI to skip the reading, so that he had time for his part time job that he needed to do to pad his resume? The purpose of higher education is to facilitate, and inspire the development of the skills and character of the student. The professor who figures out how to eliminate the use of AI on some assignments is hoping to have an impact on the both parts of that equation. But the student has to get the damn job or be destitute and burdened with thousands of dollars of debt. This will be the result without the right skills, and the most critical skill on many entry-level jobs is the ability to use AI.

This isn’t to say that the professor was wrong, either practically or philosophically. But if the university is going to accept Joe’s $50K or so payment for 4 years, we had better be damn sure that we are delivering the training that Joe needs in order to recoup that expense. 

Here is an alternate approach. As my student, use AI as much as you like. It is foolish of me to assume that you won’t. However, there will be no standardized exam. You have to explain your approach, answers, findings, and recommendations to me in your own words, on the spot. By the way, I intend to ask you questions on the spot with the same structure as the homework problems, but never the same questions.  This seems like a much better approach to the problem.