You're not the first educator to want to ban a disruptive technology. In the 1970s, professors argued that calculators would destroy students' ability to do math. In the 2000s, Wikipedia was the villain. Smartphones were going to end attention spans forever.
Sound familiar? Every generation of faculty has faced a moment where a new tool threatened to upend how students learn. And in every case, the institutions that adapted—rather than prohibited—produced better outcomes.
AI is the latest chapter. And it's a bigger one.
The Detection Problem
Many faculty members' first instinct is to turn to AI detection tools. It feels logical: if students are using AI to write their papers, use software to catch them. But the data tells a troubling story.
Multiple peer-reviewed studies have shown that current AI detection tools produce false positive rates between 1% and 12%. That means for every 100 students who submit original work, up to 12 could be falsely accused of cheating. For non-native English speakers, the false positive rate is even higher—some studies put it above 20%.
That's not a rounding error. That's a due-process problem. Several universities have already faced legal challenges after penalizing students based on unreliable detection software.
Meanwhile, students who are using AI can easily bypass detection by paraphrasing, using multiple tools, or simply running their AI-generated text through a humanizer. Detection is an arms race you can't win.
The Preparation Gap
Here's the uncomfortable truth: your students' future employers expect them to use AI. A 2025 LinkedIn survey found that 78% of job postings in knowledge-work fields now mention AI fluency as a preferred or required skill. If we send graduates into the workforce who've never thoughtfully engaged with AI, we've failed them.
This doesn't mean every assignment should involve AI. It means students need to develop judgment about when AI is appropriate, how to evaluate its output, and where its limitations lie. They can't develop that judgment if AI is simply forbidden.
What Works Instead
The most effective faculty aren't banning AI or ignoring it. They're doing something harder and more rewarding: designing instruction that engages with AI intentionally.
This means:
- Transparent policies. Tell students exactly when AI is allowed, when it isn't, and why. Ambiguity breeds confusion and dishonesty.
- Process-based assessment. When you grade the process (drafts, reflections, revisions) rather than just the product, AI becomes a tool in the learning journey rather than a shortcut past it.
- AI-responsive assignments. Some assignments should be designed to resist AI generation. Others should require students to use AI and critically evaluate its output. The key is intentional design.
- Metacognitive reflection. Ask students to document their process, including any AI use. This builds the critical thinking skills that matter far more than any single paper.
The Spectrum, Not the Switch
Banning vs. allowing isn't the real choice. The real question is: where on the spectrum does each assignment fall?
A timed, in-class essay with no devices? That's fully AI-resistant. A research project where students use AI to generate initial sources and then critically evaluate them? That's AI-integrated. Both are legitimate pedagogical choices—for different learning objectives.
The mistake is treating AI as a single binary question when it's actually a design variable you control.
Start Here
If you're ready to move beyond the ban-or-allow debate, start with one assignment. Pick something you're already planning to give this semester and ask yourself:
- What is the actual learning objective here?
- Does AI help or hinder that objective?
- How can I design this so the learning happens regardless of whether AI is used?
You don't need to overhaul your entire pedagogy overnight. You just need to start designing with AI in mind.