Should We Let College Students Cheat with ChatGPT? Why AI Cheating Exposes a Broken Grading System

AI isn’t destroying education—it’s revealing how outdated our grading systems are. A recent New York magazine investigation exposed a shocking truth: Students are using ChatGPT to write essays, solve homework, and even take exams at an unprecedented scale. But what if the real problem isn’t cheating itself, but how we measure learning? Let’s dive in.

🚨 The Real Crisis: Grading, Not Cheating

The rise of AI cheating highlights a divide in academia:

STEM vs. Liberal Arts: As the author notes, STEM fields rely on problem sets and lab work where AI offers limited utility. But liberal arts, built on essays and subjective analysis, are uniquely vulnerable to ChatGPT disruption.
1975-Scale Cheating: The author compares today’s AI cheating to 1970s marijuana use—ubiquitous, hard to police, and culturally normalized.
The Pencil Test: In STEM, exams often require handwritten problem-solving (e.g., solving differential equations), making AI irrelevant. Liberal arts assessments, however, rarely test real-time critical thinking.

✅ The Solution: Rethink Assessment, Not Restrict AI

Instead of fighting AI, educators could redesign evaluations:

✅ Project-Based Learning: Replace essays with multimedia presentations, debates, or collaborative research requiring original thought.
✅ Oral Exams: Force students to defend their ideas verbally, mimicking real-world scenarios like client pitches or policy negotiations.
✅ AI as a Tool: Teach students to ethically use ChatGPT for drafting or research, then grade their ability to refine and critique its output.

⚠️ The Challenges: Tradition, Resources, and Bias

Overhauling grading won’t be easy:

🚧 Institutional Resistance: Universities built on centuries-old lecture-and-essay models may resist change due to cost or faculty skepticism.
⚠️ Resource Gaps: Oral exams and personalized projects require more faculty time, which budget-strapped schools may lack.
🚧 The “Critical Thinking” Trap: Liberal arts programs often claim to teach abstract reasoning but rarely prove it through assessments. AI exposes this ambiguity.

🚀 Final Thoughts: Adapt or Become Obsolete

The viability of education in the AI era hinges on:

📈 Embracing Flexibility: Schools that treat ChatGPT as a collaborator, not an adversary, will better prepare students for AI-augmented workplaces.
📉 Outcome-Based Grading: Shift focus from how work is done (essays) to what students can demonstrate (problem-solving, creativity).
🚀 STEM Lessons for All: Adopt STEM’s emphasis on applied skills—like labs and exams—across disciplines.

So, what do you think? Is AI cheating a wake-up call for education—or a threat to academic integrity? Let’s debate in the comments.

Let us know on X (Former Twitter)

Sources: Jonathan V. Last. Unpopular Opinion: Let College Students Cheat with ChatGPT, May 24, 2025. https://www.thebulwark.com/p/unpopular-opinion-let-college-students