Beyond AI Detection: Designing Courses That Make Student Thinking Visible
The Brown University AI cheating case highlights the limits of AI detection. Learn why better course design, active reading, and social annotation offer a more sustainable path forward.
When news broke that a Brown University professor suspected many of his students had used generative AI on a take home midterm, the story sparked a familiar conversation across higher education.
Should instructors return to handwritten exams? Should universities invest in AI detection software?
How should institutions respond when AI changes the way students complete assignments?
You can read the full story in Inside Higher Ed: Brown Professor Suspects Majority of His Class Used AI to Cheat
These are important questions.
But buried within the article is an even more important idea: The future of academic integrity may depend less on catching AI after the fact and more on designing learning experiences where authentic thinking happens throughout the process.
AI Detection Has Limits
One of the most revealing parts of the Brown story came not from the professor, but from Brown University’s own committee on generative AI.
Rather than recommending stricter detection methods, the committee acknowledged a reality that many educators have already experienced:
“There is no way to check with 100 percent accuracy whether GenAI has been employed.”
Brown University Committee on Generative AI
Instead of relying solely on punishment, the committee encouraged faculty to rethink assignments, clarify expectations, and create opportunities for open conversations about AI use.
That recommendation reflects a broader shift happening across higher education.
Many instructors are realizing that detection alone cannot solve a teaching problem.
The Better Question
For years, conversations about AI in education have centered on one question: Did students use AI?
But there is another question that may be even more valuable: What did students do before they used AI?
Did they engage with the reading? Did they identify the author’s argument? Did they ask questions? Did they wrestle with difficult ideas? Did they build their own interpretation before asking AI for help?
These are the kinds of learning experiences that develop critical thinking, regardless of whether AI is eventually part of the workflow.
Better Learning Starts Earlier
Traditional assignments often reveal very little about how students arrived at their final answer.
An essay, discussion post, or exam shows the finished product, but not the thinking behind it.
By the time instructors review student work, the learning process is already complete.
That creates a challenge in the age of AI.
If instructors only see the final submission, they miss the questions students asked, the evidence they considered, the misconceptions they encountered, and the moments where learning actually happened.
Make Reading Part of the Assessment
One way to redesign assignments is to place greater emphasis on the reading process itself.
Instead of treating reading as preparation for learning, instructors can make reading part of the learning experience.
Social annotation allows students to engage directly with course materials by highlighting passages, asking questions, responding to classmates, identifying evidence, and building interpretations alongside the text.
The result is that learning becomes visible long before a final assignment is submitted.
Faculty gain insight into how students are thinking, not just what they ultimately write.
Design for Independent Thinking First
One concern many instructors have is that students may simply repeat one another’s ideas.
That is why instructional design matters.
With Paced Social Annotation, students first complete their own annotations independently before viewing their classmates’ contributions. This encourages students to develop their own interpretations before collaborative discussion begins.
Rather than replacing collaboration, pacing strengthens it by ensuring every student has an opportunity to think first.
Learn more about Paced Social Annotation →
More Engaging Assignments Create Better Conversations
One of the most encouraging aspects of the Brown report is that it shifts the conversation away from punishment alone.
The committee recommends helping faculty adapt their teaching practices rather than simply enforcing new restrictions. That is where assignment design becomes especially powerful.
Instead of asking students only to produce a final essay, instructors can invite them to document their thinking throughout the learning process.
They might identify an author’s central claim.
Question an assumption. Highlight supporting evidence. Respond to a classmate’s interpretation. Connect one reading to another.
These activities are difficult to outsource because they are rooted in each student’s engagement with the course material.
Our Generative AI and Social Annotation case study explores how instructors are using annotation to create more authentic learning experiences in the age of AI.
Build Learning Around Primary Sources
As AI becomes more common, helping students engage with primary sources becomes even more important.
Social annotation supports active reading across journal articles, websites, videos, images, Open Educational Resources, publisher content, and many other types of instructional materials.
When students actively interact with original sources before generating a final product, instructors gain a much clearer picture of their understanding.
You can also explore our Complete Guide to Social Annotation for Online Reading for practical strategies and assignment ideas.
The Future of Academic Integrity Is More Engaging Learning Design
Artificial intelligence is changing higher education, we are all aware of that. What remains within educators’ control is how learning experiences are designed. The Brown story reminds us that AI detection will never be perfect.
Fortunately, it does not have to be.
When instructors create opportunities for students to engage deeply with readings, ask meaningful questions, collaborate thoughtfully, and make their thinking visible throughout the learning process, academic integrity becomes about more than preventing misconduct.
It becomes about cultivating genuine learning. That is a goal worth designing for.
RELATED RESOURCES
- Paced Social Annotation
- Generative AI and Social Annotation Case Study
- The Complete Guide to Social Annotation for Online Reading
- Brown Professor Suspects Majority of His Class Used AI to Cheat
Ready to Design Learning for the Age of AI?
The most effective response to AI is not simply detecting its use. It is creating learning experiences that encourage students to read deeply, think independently, and engage meaningfully before producing their final work.
Hypothesis helps instructors make learning visible through collaborative annotation, giving students more opportunities to develop critical thinking while giving faculty greater insight into the learning process.
See how Hypothesis can help your institution design more engaging learning experiences →