AI Menace: How Fake Students and Bots are Infiltrating Online College Classrooms
These are not the droids you are looking for - Ben Kenobi
Like an army of ants invading a picnic, bots are online everywhere, even in my community college classrooms. This is one of the many reasons it has become increasingly difficult to separate reality from fiction online. Bots have started popping up in the direct message chat here on Substack with creepy small talk about the weather. Even more troubling, Meta is flooding social media with AI influencers. Being online has begun to feel empty and isolating, reinforcing the idea of the dead internet theory.
Fraudulent students using AI bots to infiltrate online classrooms have me and many of my faculty colleagues questioning who is authentic, who is an actual person, and who these bots are. I long for the tangible, the recognizable students, some of whom are still struggling with lost learning from the COVID-19 times. Some students have heartbreaking needs; some are struggling to find a clean and safe place to sleep, others are trying to manage food insecurity, and more.
But faculty time is spent trying to figure out who is real and who isn’t. In recent semesters, about one-third of the students in some of my classes have been fraudulent.
And not just me and my classes, or even only my college. From November 2024, an article by EdSource explains,
California Community Colleges have lost more than $7.5 million to sophisticated online scammers this year.
In June of 2023, the San Francisco Chronicle also explained,
Today, about 20% of California’s community college applications are scams: more than 460,000 of the 2.3 million requests to the state’s online application system since July alone, says the state Chancellor’s Office, which oversees the 116 campuses.
EdSource explains further,
Colleges have increased their efforts to detect and deter fraud through more human interaction and automated detection. Officials believe they are getting better at doing so, but the increasing losses show that the college system is still vulnerable to scammers, who are often part of sophisticated crime rings, some overseas.
In an effort to help more students and provide instruction that better serves Community College students with their diverse needs, Community colleges have long been susceptible to fraud, since they are generally open access and usually don’t deny admission to students who meet basic requirements as the more selective University of California and California State University do. And this problem (like so many others) was made worse by the Covid-19 pandemic. The shift to remote instruction, a lifesaver for so many during a chaotic and turbulent time, “created fertile ground” for fraudsters, said Paul Feist, a spokesperson for the chancellor’s office overseeing California’s community colleges. The scammers wanted to get their hands on the nearly $2 billion in federal stimulus dollars available for emergency student aid available across the colleges.
These scams are frustrating for instructors like myself and students who need a seat in a course that a fraudster holds. With a finite number of seats for each course, real students are often left on waiting lists and unable to enroll in necessary classes because fraudsters are taking up that space.
In addition to taking money from institutions and seats from actual students, combatting fraud has become a never-ending battle as fraudsters evolve and develop new tactics. In practical terms, having fraudulent students in my online classes has changed how I do many things.
For example, I have a check-in assignment for online classes that is not a simple introduce yourself text-based discussion post like I used to do. Now, I have students record an intro video to check into the class. Video check-ins are more difficult for AI to complete.
But remember, I said they’re evolving. This past semester, I got my first AI student intro video. Of course, I can drop the fraudulent student, but often, as soon as I drop one, another fake student on the waiting list takes their place. It can feel like an endless few weeks of whack-a-mole for fraud students, and it takes up an enormous amount of time that should be devoted to real people and building a welcoming learning community.
One of my courses last semester had 30 students enrolled to capacity, but by the end of the first week, I had culled the bots, and it was down to just five students. I cannot know if students who wanted the course were turned away, which is also frustrating. As a result, colleges can also be fooled into adding courses to accommodate a perceived higher need.
These fraudulent students use fake avatars and photos. I have done reverse image searches to find out they are stock photos. There are other idiomatic distinctions, at least for now. Bot names are sometimes in all capital letters or have a repeated name within their name, like Robert Robert Smith. These are the easy-to-spot versions of the bots. My fraudulent student sniffer is pretty strong, but getting used to spotting them takes a bit of practice. I didn’t go to school for decades to become an expert on sniffing out fake students in my classes, but here we are.
Cal Matters has a name for the fraudsters,
They’re called “Pell runners” — after enrolling at a community college, they apply for a federal Pell grant, collect as much as $7,400, then vanish.
Ed Tech magazine calls them “ghost students.”
It is also a tricky balancing act to manage because sometimes, real students get kicked out of classes, which is something I don’t want to happen. Worse, clamping down too hard on fraud can have other unintended consequences. More than 20% of community college students in California don’t receive the Pell Grants for which they’re eligible. Many of these students need this money to get through college. Also, California taxpayers are paying for them when this money should be helping educate the next generation.
I wish I could say that surviving the first few weeks of bot whack-a-mole would mean the remaining students in every online class are actual students, but some bots still sneak through. If they turn in assignments, they are often gibberish, although a small handful submit gradable work. Another waste of time is instructors having to grade fraudulent student work.
This phenomenon has forced instructors to focus on hunting bots instead of teaching. Instructors are now expected to test their students in the first weeks, asking them to submit answers to prompts, sign copies of the syllabus, or send other evidence to prove they are real.
To make matters even worse, adjunct instructors are being paid per course and make substantively lower wages than the full-time faculty for teaching these courses, so this added burden on their workload is exploitative to them as instructors.
Is anyone doing anything to fix it?
According to Cal Matters,
The largest reform underway is a new version of CCCApply, the state’s community college application portal, which will offer more cybersecurity, Feist said. He also said there are other “promising” short-term projects.
One of them, a software tool known as ID.Me, launched in February. The contract with the software company, costing more than $3.5 million, gives it permission to check college applicants for identification, including video interviews in certain cases. Privacy experts have warned that the company’s video technology could be racially biased and error-prone, nevermind what this means in regards to student privacy.
Additionally, the California Community Colleges academic senate released guidelines in 2021 to help faculty engage in ways that can mitigate enrollment fraud. The guidelines include engagement strategies to use in the classroom and some practical and logistical steps to detect and report fraud, much like the ones I mentioned that I am currently using in my classes.
My institution's administration has been doing everything it can to help combat the issue, including having students make same-day tuition payments and jump through other enrollment hoops to ensure real students get into classes. However, many of the fixes employed by community college leadership and local college administration to mitigate bots end up being barriers to actual students, which is undesirable.
The story of fraud in community colleges isn’t getting much coverage in the mainstream media sphere, which is also problematic. I found this story's information primarily in niche education-based outlets like Ed Source, Inside Higher Education, and EdTech Magazine.
The most mainstream coverage I found was an article in Forbes and the San Francisco Chronicle. The best coverage has come from the non-profit Cal Matters, which has been updating its readers on the issue since the pandemic.
Teaching is very rewarding. I learn so much from my students and have made lifelong friends with past students. But all of this bot-hunting and whack-a-mole can suck the joy right out. It definitely makes for a challenging and cumbersome first few weeks every semester.
I just looked at my classes this semester and can already tell I have a bot problem with some. So, this is still an ongoing issue.
Wish me luck. Classes start January 13th.
Good luck with the whack-a-mole! Gonna be a fun few weeks.
Your bot sniffer examples are actually super helpful info, for which I thank you!!!
Now, this is quite interesting... 🤔 It's definitely an eye-opener.