Cheating on essays and academic papers has existed for centuries. From plagiarism to Wikipedia, too many students have always found ways to avoid the hard work of writing. But today, a new tool has rapidly become the preferred shortcut: Large Language Models (LLMs) like ChatGPT, Gemini, and Grok. Hence, the desperate need for AI detectors.
Unlike traditional plagiarism, LLM-generated content often evades detection. It doesn’t directly copy existing material, but instead generates seemingly unique text—making plagiarism checkers ineffective. With a little prompt engineering, students can produce unlimited essays that appear “original.”
Educators Are Giving Up Too Soon
A troubling trend is emerging: many educators are surrendering to AI, saying things like, “We need to prepare students to be AI-proficient or they’ll be left behind.” While learning new digital tools is important, AI “proficiency” is a pretty low bar. Worse, way too many educators, including writing instructors, are no longer demanding that students learn by producing original work.
AI Hurts Your Brain
Research from MIT demonstrates that outsourcing thinking to AI can diminish cognitive development. You’re not going to build muscle by watching someone else lift weights, students don’t become smarter by letting LLMs do most of the heavy lifting. We learn by doing—reading, solving, writing—not by automating those processes before the students have developed mastery of the subjects.
Remember the Calculator Debate?
Probably not, but I was around when calculators first became widespread. This debate is similar to when calculators first entered classrooms. The solution then was balance: once a student learned the fundamentals like addition, subtraction, multiplication, and division, calculators helped them tackle more complex math problems. Students spent their time solving these complex problems rather than on rote math problems. Writing should evolve similarly. We must continue teaching how to write, even as we leverage AI for brainstorming, editing, and refining content, but not replacing foundational learning.
Can AI-Generated Content Be Detected?
Yes, but we there is an “arms race.” AI detection tools, often called AI text classifiers, look for statistical patterns found in machine-generated language. LLMs often overuse phrases that a human normally would not. “Humanizers” rewrite AI output to bypass detectors, often at the expense of coherence and sounding “like a human.” The best detectors, like those from Pangram Labs, are catching on by identifying these unnatural patterns and attempts by humanizers to mask those unnatural patterns.
Many early AI detectors are themselves powered by LLMs which worked in the beginning, but are now increasingly unreliable. These products too often yield false negatives (missed AI usage) and false positives (misidentifying human work), creating serious trust issues in academic and professional contexts.
The Double-Sided Business of Detection
Too many AI detection companies now also sell humanizers. They profit from helping students cheat and profit again by helping educators stop it. A business model analogous to selling weapons to both sides of a war with the purpose of escalating the war so they can profit even more. Since there's often more money on the cheating side, AI detectors must continually outsmart the humanizers.
Pangram Labs has taken a different route, investing in proprietary AI models built specifically for detection and constantly updating them to keep up with new LLMs and humanizer tools. They've recently launched a plagiarism detection feature, further expanding their platform to help protect academic and intellectual integrity.
Final Thought
Cheating isn’t just an ethical issue it’s a cognitive and reputational one as well. In a world where generative AI is pervasive, preserving the integrity of learning, creativity, and original thinking has never been more important.
Don’t give up on teaching the fundamentals. We need to evolve how we teach, but not abandon the importance of teaching in the first place. AI detectors keep us honest and smarter.