AI-Proof Project-Based Assessments by Making Them Context-Specific
Project-based assessments that are rooted in real-world, institution-specific contexts can reduce over-reliance on generative AI tools. They do this by emphasizing collaboration, oral defence, and critical thinking. Here’s how to design such assessments effectively.
Why Choose Project-Based Learning?
Generative AI’s ability to produce content quickly poses a challenge for language teaching, especially when assessing writing skills. At many institutions, students increasingly depend on AI to complete writing assignments, which weakens the value of traditional essays as a measure of language proficiency.
One way to address this is to modify assessment tasks to be more creative and context-specific. Project-based learning (PBL) offers a strong alternative. It shifts the focus to student-driven collaboration, problem-solving, investigation, and reflection within real-world contexts.
For example, students might work together to propose improvements to their campus environment. They could suggest installing beehives and flower gardens to boost biodiversity and enhance the campus atmosphere. Early in the project, students explore and document the current state of the campus, considering potential locations for their proposals.
They gather data through surveys of students and staff, discovering concerns such as allergies or phobias related to bees. Throughout the semester, instructors guide students as they brainstorm, critically assess ideas, and develop a collaborative written proposal, such as a 2,000-word document outlining their plan.
This approach positions teachers as facilitators and students as active problem-solvers addressing institution-specific challenges that AI cannot fully grasp.
Adapting Assessment Criteria
Since AI tools can polish written text, linguistic accuracy becomes less useful as a primary evaluation metric. Instead, project-based assessments can reduce the weight of the written component and increase the importance of an individual oral defence.
Start by assessing how well students work together to produce a logically sound and coherent written proposal. Evaluate their persuasive communication, such as how convincingly they use survey data and secondary sources to support their ideas.
Then, conduct an individual speaking test where students defend their proposals in response to examiner questions. These questions might be broad, like “What challenges did you face during the project?” or specific, such as “Why did you choose that location for the beehives?”
Using Common European Framework of Reference for Languages (CEFR)-aligned criteria, assess students on their ability to articulate ideas clearly and convincingly in real time. This balances the focus between writing and oral communication skills.
Rewarding Contextual Awareness
Because the task requires specific knowledge about the institution and academic style, AI tools cannot effectively complete it. For instance, both students and examiners know that beehives can’t be placed on the third floor of the central building, but AI might not.
By designing assessments that involve tangible, context-specific issues, the temptation to rely on AI for writing tasks diminishes. This shift encourages students to engage directly with their learning environment rather than outsourcing their work to AI.
Importantly, this approach doesn’t punish AI use; it simply makes it less effective for completing the task. Compared to timed exams, project-based assessments offer a positive way to incorporate AI thoughtfully into higher education.
Keeping assessments flexible and responsive ensures they remain effective tools for measuring genuine language proficiency, even as AI technology evolves.
Your membership also unlocks:
