One-Size-Fits-All AI Doesn’t Fit the Classroom

AI is in classrooms and lecture halls around the world, whether we’re ready for it or not. The question now is: When and how should AI be used, and when should it not, to protect the development of critical thinking skills?

At Turnitin, we work with more than 16,000 institutions around the world and have been around for nearly 30 years. We have seen our fair share of technological shifts, but nothing has upended education like AI. Now that its use is mainstream in most countries, education faces the challenge of whether and how to use it (knowing that many students already are). There isn’t a one-size-fits-all answer, so it’s important to stay up to date with the latest insights on AI use and trends to help inform administrator, educator, and student decisions around AI. Here’s what I’m seeing right now:

Students are Using (and Misusing) AI

The latest Learning Integrity Insights Report analyzed millions of submissions from around the world and found that AI use in writing is very common. Amongst US higher education submissions, about half (49.9%) of submissions have at least some AI writing detected, and nearly 1 in 5 (19.4%) had an 80% or more AI writing score, reflecting a near-total outsourcing of their writing to AI. That’s nearly double the rate of their peers in the UK (9.8%) and Australia (10.2%) with greater than 80% AI writing scores.

In secondary education, AI use is lower, finding less than a third of US submissions (29.8%) with any AI writing detected and 6.3% with more than 80% AI writing. This level of use in the US reflects a tipping point, emphasizing the need for clear guidance around AI use for both students and educators and the transparency needed to protect critical thinking.

Students are Hungry for Feedback From AI

To better understand how students are interacting with AI assistants, the report analyzed students’ assignment prompts in our formative writing product called Clarity, and found them hungry for feedback. Students were most likely to ask a question when they were in the drafting stage, defined as an active writing phase when students are developing the content of the paper.

The most common request was to ask for feedback on the content of the paper, such as: “Is my paper good?” or “Am I aligned with the rubric?” or “How can I improve my paper?”. Requests for feedback about structure, such as paragraph organization and use of transitions, and style, such as how to format citations, were the next most common.

As a parent, I love the idea of a real time writing coach that is designed to not write for the student, only guide their development as writers and thinkers.  This is an example of “AI on guardrails”, and I believe this is how AI should be deployed in education.

Educators (Not University Leadership) are Leading on AI Policy

There is clearly a spectrum of openness to AI use in student work, across faculty, and even across individual assignments. Clearly there is a real focus on preserving learning integrity, while also building AI literacy. Forty percent of poll participants reported teaching, learning, and faculty leaders were primarily responsible for setting AI use strategy, compared to 24% reporting executive and academic leadership and 18% sharing that a cross-functional committee drove the strategy at their institution.

Nearly half (48%) of those polled reported that teaching and learning leaders and faculty were primarily responsible for implementing AI-related tools at their institutions, with only 17% saying IT and technology leadership and 16% sharing that cross-functional committees were responsible. In the past, we would have expected more IT departments or other administrators involved in technology strategy and implementation.

I think this shift reflects a growing recognition that successful, responsible AI adoption depends on educators having visibility into how students are using AI, along with the flexibility to incorporate these tools in ways that align with their instructional goals.

If there’s one thing I’ve noticed recently, it’s the need for flexibility with AI in education. I don’t believe it’s appropriate all of the time, or none of the time. Educators need to be able to dial it up or down, and institutions need to set clear guidance on what’s permitted, such as outline help, and what’s not, such as use on a high-stakes exam.

Education requires a spectrum of approaches to better support the development of critical thinking skills. Students, educators, institutions–and even parents and employers–benefit from clear expectations around AI use that are rooted in learning integrity rather than simply introducing new technology.