AI is Our Steroid Cheating Crisis
Plus, one more cheating technology to worry about. Plus, University of Memphis sits a basketball player for alleged cheating.
Issue 278
To join the 3,767 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 17 amazing people who are chipping in a few bucks via Patreon. Or joining the 32 outstanding citizens who are now paid subscribers. Thank you!
AI is Steroids. We Need a Solution.
Noor Akbari, founder of integrity company Rosalyn.ai, wrote a must-read column on AI and academic integrity in Ed Week.
I cannot recommend it enough.
In it, Akbari likens the use of AI in education to the use of steroids in competitive sports, and he is bullseye accurate. Both make it impossible to tell what someone did, what they’re actually capable of doing. Both make a mockery of fair competition and make assessment against any kind of norm impossible. Moreover, as he points out, both are dangerous — one to the health of the user, the other to safety, equity, and society.
Were it possible to see someone stand in ovation in a newsletter, you would be doing so now. Please go read it. Please go share it.
Among other important points, Akbari says:
At the height of the sports doping crisis, a common argument was to let doping happen. If no substance is banned, isn’t the playing field level? The counterargument is that substances are banned because they pose a health risk to athletes. If sporting organizations not only allowed but tacitly encouraged athletes to dope, the resulting biochemical arms race would have a sure loser: athletes and their well-being.
Likewise, the normalization of AI doping would create an arms race among students, resulting in several consequences for them and society
Yes.
A “just let everyone use it” approach is wildly foolish because, he says, it undermines a fundamental value of education — actually learning how to do things. Akbari says such an approach will put unqualified people in positions to do terrible things. And, in a argument we do not hear nearly often enough, he says a universal acquiescence to AI in education could:
produce a two-tiered education system that is even more inequitable than the current one. Elite colleges with full-time professors and graduate students have the resources to design assignments in which AI provides no edge. Community colleges and online education platforms don’t have that luxury. A class with hundreds of students and one part-time instructor cannot convert every digital test into an original research project or in-person test with pencils and paper—not without raising costs considerably.
Efforts to democratize education will be laughable if the only credible degrees come from private, in-person institutions that cost students an average of almost $56,000 per year.
Exactly. The market already understands the difference between a degree from Yale and one from Southern New Hampshire University. It’s not just reputation, it’s reputation based on rigor. And once the market internalizes which types of schools allow unfettered use of AI steroids, the value of certifications from those schools will change. By change, I mean crater.
Just two more things to share from this piece. One:
Forget watermarking AI-generated text and AI detectors—they’re easily duped. And forget academic “honor codes.”
I am thankful he did not say that AI detectors do not work. I think he knows they do. But he is right that they, like all integrity and security systems, can be bypassed. That does not mean we should not use them. As I say all the time, it’s not too hard for someone to break a door lock. No one suggests not locking your doors.
He’s right about the “honor codes.” Recent research is pretty clear that they are pretty useless.
Finally — but again, please go read it — Akbari says:
The silver lining of the 1998 Tour de France was that the International Olympic Committee formed the World Anti-Doping Agency the following year. Though far from perfect, WADA created a unified list of banned substances and standards for detecting them. In other words, the organization defined what “doping” means in sports. Soon enough, education systems may need a WADA-like organization to define cheating in the AI age and set standards for preventing and detecting it.
Doping in sports undermined the fairness and meritocracy of a beloved institution, until that institution took the threat seriously. It’s time we take AI doping in schools seriously.
I agree entirely.
Unfortunately, I am skeptical that any such entity will surface. So far, despite overwhelming evidence of the scope and consequences of academic misconduct, no one seems motivated to do much about it — not government, not accreditors, not institutions, and not most educators.
Most likely, we will all just continue to look away and be amazed at how much faster and stronger someone can be if they use steroids.
Just One More Thing to Worry About
Cheating and cheating prevention are often described as an arms race. Well, according to this post and video on LinkedIn, the cheating side has another weapon. At least potentially.
If you cannot access the post, or care not to, it’s about an AI wearable device that will listen to what you say or ask and laser-project an AI-generated answer on the palm of your hand. Nice, right?
So, test centers and exam proctors beware. As these devices become available, any student could repeat any exam question aloud and simply look at their hands, in the direction of a test or keyboard, and have the answer projected for them. The demo in the LinkedIn clip is insightful.
And just wait until these get smaller and less obvious. So, one more thing. Sorry.
University of Memphis Basketball Team Benches Player over Academic Misconduct Inquiry
The University of Memphis basketball program is a big deal and, according to news coverage, the team has decided not to play a significant player due to allegations of academic misconduct.
Details of the allegations were not reported.
I love this for two reasons. One, good for the athletic program and its leaders for modeling good action. Cheating has consequences and, at least until a determination can be reached, better to not play. I feel as though it’s a strong message to send.
Second, probably a major reason for the decision to sit the player is that the NCAA — the governing body for college athletics — has punished teams for academic misconduct violations in the past (see Issue 130 or Issue 136 or Issue 172 - and there are others). This means that, had the school known about the alleged cheating — which they did — and not taken action, there could have been serious consequences.
The potential of more significant consequences from the NCAA shows me that schools or (some) educators are unlikely to take academic cheating seriously until someone makes them. In sports, cheating — even academic cheating — can result in suspensions, invalidation of records and awards, and fines. Nothing even remotely like that exists outside of athletics. It should.