Brandon University introduces communication tool for student use of GenAI in courses 

Three-level system provides instructors flexibility on GenAI use in courses

Instructors at Brandon University (BU) are adopting a new system that gives them the ability to effectively communicate whether they will accept the use of Generative Artificial Intelligence (GenAI) in their courses. 

Introduced in May 2024, the system which uses the analogy of a traffic lights to describe how much, if any, GenAI is allowed in a particular course.  

The system is a result of extensive research into best approaches for navigating the new world of GenAI, which has made contract cheating less attractive with its ability to create low-level essays. Curt Shoultz, Director of BU’s Centre for Teaching, Learning, and Technology (CTLT) led the development of the policy, which has been received favourably by BU instructors.  

“It made no sense to have a one-size-fits-all structure” Shoultz says. “So, the rules are determined by each individual professor for each course – because there may be room for GenAI in some courses and not in others.”  

In CTLT the suggested system the traffic light colours stand for the following acceptable GenAI use levels: 

• • Red light: No GenAI is allowed in the course, other than what is available in word processing software.  

• • Yellow light: Some GenAI is allowed, within the parameters provided by the instructor and within the university’s academic integrity guidelines. 

• • Green light: GenAI is allowed in the course, within academic integrity guidelines. 

In each case, it is critically important that instructors have a good conversation with their students about why or why not GenAI is allowed in the course, Shoultz explains. Students need to understand the intent of its use and how it might either interfere with or enhance their learning experience. 

For example, instructors who do not allow any use of GenAI may explain that they want their students to generate their own ideas, edit their own work themselves and be thinking critically at every level of the essay, from word choice to sentence construction to overall organization.  

Instructors can give their course a yellow light for a variety of student scenarios. “An instructor might say students can use GenAI to break the blank page, if they are really anxious about starting a first draft.” says Shoultz. Or the reverse might be allowed, where the student creates a first draft and then uses a program like Grammarly to improve upon it.  

If a course is given a green light for GenAI, it might be because there is value in exploring the use of GenAI as part of the discovery process of learning. It might be needed to prepare students for future roles, research areas or industries.  

When a student uses GenAI in a course, they ideally would track the changes in their document to show where GenAI has been employed. They might also be asked to include comments to explain choices they have made on its use.  

What about trying to detect the use of impermissible GenAI in student work? The use of electronic detection is discouraged, says Shoultz. 

“We haven’t seen anything yet that can detect AI 100 per cent of the time,” he says Shoultz. “Using electronic detection methods can create problems with both false positives and negatives and accusations of plagiarism that are unverifiable.”   

Instructors are encouraged to employ the following techniques: 

• • Have conversations in class early and often about why they are accepting or not accepting the use of GenAI. If students understand the reasoning, most will understand that the rules are in place to help them learn.  

• • Become familiar with their students’ work. Early in the term, having students write something brief in class can provide a writing sample to compare their future work to.  

• • Have confidence in their ability to detect most instances of GenAI, which is often characterized by verbosity and vagueness.  

• • Avoid having just a few high-stakes assignments for the class. The higher the stakes, the greater the incentive to cheat.  

“We’ve had a lot of people say they like the system,” says Shoultz. “We synthesized a lot of the current research to offer the best suggestions for what to try.” The traffic light system will be updated or redesigned as need as the effects of GenAI in assessment are better understood in post-secondary context. 

“We’ll keep talking about what’s working and what’s not.”