Creating assessments in the age of AI
In light of Artificial Intelligence advancements and the ever-increasing popularity of AI tools and LLMs (Large Language Models), many educators recognize the need to reevaluate the way we assess learning. We need to think critically about the design of not only individual prompts but the overall structure of our assessments. Here are a few ideas to consider, as we seek to develop assessments that discourage students from relying on AI-generated responses as a substitute for their work and demonstration of knowledge
Suggestions for assessment design
- Many of our AI-specific concerns lead us back to good, solid pedagogical practices. While certainly not a new idea, focusing on iterative writing processes and scaffolding assessments into smaller parts is especially important in this context. By reducing the anxiety students face when confronted with large, overwhelming summative assessments, we also reduce the temptation to cheat or use AI inappropriately.
- For assessments that require students to produce written responses, consider adding a self-reflection component. For example, if students are writing 10-12 page term papers, they might also be required to complete a reflection activity asking them to reflect on the actual process and experience of writing the paper. This reflection might be in a different format, such as a short video where students are asked to candidly share and reflect on their writing process. The benefits of reflective writing are well-documented, as outlined on this site from Purdue University. This resource from George Mason shares some helpful ideas for reflection questions.
- Experiment with AI together with your students and analyze the results. Use AI as a tool to demonstrate good writing and editing skills. For example, at the start of the semester, you might ask students to respond to a prompt in class. Split your class into three groups; one group uses only AI to complete the assignment, one group uses AI and then edits their response, and the last group completes the prompt without AI. Come back together as a large group to analyze the results and reflect on the writing process.
- Incorporate in-class writing time. For those who teach in person, dedicating more in-class time to writing in real time may be an option. Think about opportunities for low-stakes, formative writing assignments. This will not only help build students’ skills and confidence with writing but can also provide you with a better understanding of their baseline writing abilities for comparison if inappropriate AI use becomes a concern.
- Offer suggestions for appropriate AI use. Although concerns about the inappropriate use of AI are well-founded, it’s also true that AI can be a useful tool for learning. Feel free to offer students suggestions and examples of what appropriate use of AI might look like for a particular assignment. Perhaps they might use it to brainstorm titles for a project, or prepare a timeline for completion, etc. Modeling appropriate use of AI helps students not only in your class but also allows them to develop skills that might assist them in their futures.
Additional resources
We hope these suggestions offer some helpful ideas for reevaluating how we can assess our learners effectively and authentically. More ideas from our colleagues at OTEAR, as a result of the AI Roundtable Initiative, are available for your consideration here. If you have any questions or feedback, please feel free to contact us at idt@camden.rutgers.edu.