A Closer Look at AI Detection Tools’ Footprint in Higher Ed

Aug 11

By Claire Brady, EdD

A note about today’s blog: I don’t work for ed tech companies—I work with institutions, helping them navigate their options, evaluate tools, and make strategic decisions. I’m not neutral, but I am agnostic: I care less about which vendor you choose and more about whether it truly serves your mission, your students, and your values. And when it comes to AI detection tools, we have to be honest—they’re not yet as advanced or reliable as we need them to be.

As higher education continues to grapple with the impact of generative AI, many institutions have doubled down on surveillance-based solutions like Turnitin to preserve academic integrity. But a powerful new investigation by CalMatters and The Markup—“California colleges spend millions to catch plagiarism and AI. Is the faulty tech worth it?”—raises urgent questions about the financial, ethical, and educational consequences of this approach.

For higher ed leaders, this article raises an important question: Are tools like AI detectors—despite their cost—reliable enough, and are we fully considering the potential risks they pose to student privacy and trust?

The Cost of Comfort

California’s public colleges and universities have collectively spent over $15 million on Turnitin tools, despite mounting evidence that AI detectors produce false positives and fail to catch actual AI-generated writing. Students who never used ChatGPT are flagged. Native English speakers breeze through while multilingual learners and students who rely on grammar tools like Grammarly are disproportionately penalized. And yet, institutions keep renewing these contracts—largely out of perceived necessity and lack of options in the marketplace.

Rethinking Our Priorities: Support Over Surveillance

One of the most pressing takeaways from the article is the misalignment between institutional investments and what students and faculty express that they need. Instead of funneling resources into AI detection tools with questionable accuracy, colleges could focus on equipping their communities—through AI literacy training, clear academic integrity policies, and stronger student-faculty relationships. The investigation also revealed troubling equity issues as well. Black students and non-native English speakers are more likely to be falsely flagged, caught in the crossfire of imperfect algorithms.

What Higher Ed Leaders Can Do Now

Audit and Question Existing Tools: Does Turnitin or your AI detection software actually work? Are the risks to student data and trust worth the return?
Invest in People, Not Just Products: Prioritize faculty development and student guidance over automated policing. AI is here to stay—our response must include education, not just enforcement.
Build Ethical Frameworks: Create policies that define acceptable AI use, protect student work, and emphasize transparency and fairness.
Reimagine Assessment: Move beyond essays that can be easily completed by a chatbot. Incorporate oral exams, reflections, collaborative projects, and in-class writing.
Lead with Trust: As the article powerfully reminds us, deterrence doesn’t build integrity—relationships do.

In the end, AI detection tools may offer institutions a sense of control—but at what cost? The answer, as this investigation suggests, may not be in the software—but in the values we choose to center on our campuses.

Turnitin’s Reach

Accuracy & Effectiveness Concerns

Turnitin’s AI detector does not detect hallucinations (fake quotes, made-up sources)—a key signal of AI-generated content.
The software often flags students who didn’t use AI, especially those using Grammarly or other editing tools.
The AI detector has been fine-tuned to reduce bias but still falsely flags non-native English speakers and students with simpler syntax more often.

Privacy & Ownership Issues

Turnitin holds a “perpetual, irrevocable, royalty-free license” to all submitted student writing.
Its database contains 1.9 billion student papers as of June 2025.
Unlike previous vendors (e.g., VeriCite), Turnitin does not allow students to opt out of having their papers stored in the global database.

Student Impact & Equity Concerns

1 in 5 high school students surveyed said they or someone they know had been wrongly accused of using AI to cheat.
Black teens were twice as likely as white or Latino teens to be flagged for AI-generated work—according to Common Sense Media.

Read the full article here: https://calmatters.org/education/2025/06/california-colleges-plagiarism-turnitin/