10 Tips for AI in Assessment
Ten ideas to keep in mind when assessing students in an AI world
Table of Contents
Estimated time to read this article: 7 minutes
What?
Some form of artificial intelligence (AI) is already used regularly in our daily lives and in our academic work. Grammarly, predictive text, citation machines, and homework help sites are all AI based tools that we use to complete our work. Siri, Alexa, and Google Assistant help out around the house. Social media, online recommendations and navigation tools also include elements of AI.
Designing assessment in this new AI driven world is challenging but invigorating. Here is an opportunity for us as faculty to change up (even spice up) our assessments to push our students to reach our course learning outcomes. Below are a series of ideas and suggestions to help with the design of an assessment that remains relevant in an AI world.
So What?
Start with answering “What kinds of assessments do you use in your class?”
Now, it's time to explore and brainstorm. At the end of each numbered section, stop and reflect. Record your thoughts and ideas in a manner that works for you.
- Explore how AI handles your subject area
- How does AI handle your assessment prompts?
- Try different AI tools (ChatGPT, Bard, Co-Pilot) - different tools have different strengths.
- Adjust prompts as systems evolve.
- Decide what will work best for you, your content and your students.
- Think authentically and alternatively
- Allow for creativity and alternative means of expression.
- Match depth of assessment expectations with depth of learning outcomes (e.g. Blooms taxonomy).
- Flex assessment format, unless specified in learning outcomes.
- Consider: Experiential learning or community service learning, Hands on assignments, class discussion, peer review, Personal reflection/opinion based on course material (e.g. Apply theory to personal experience), Custom course materials (create an assignment based on), In-person or video presentations/oral Q&A or reports, real world projects (authentic assessments)
- Ask for critical, local, and personal thinking
- Focus on critical analysis of gathered data, less on introduction and theory components
- Emphasize requirements of experience and personal thought and analysis rather than use of sources and citations
- Instead of: "write a self-reflection"… try "How does this content connect to your experiences" or instead of "What is …?" … try "Based on what you learned in class, how would you …"? or "Make a prediction what would happen if ..."
“Assessment of understanding should be thought of in terms of a collection of evidence over time instead of an “event” – a single moment-in-time at the end of instruction.” (Auburn Online, 2024).
- Rely on our library and community
- Require use of specific resources from our library
- Require communication with community members/ organizations to support learning
- Teach and demo citing use of AI
- Think career wise and not academic wise
- Consider multimodal assessments:
- Lab report: use video presentation of procedure and results with annotated diagrams
- Analysis paper: use digital collage with text, images, videos
- Research paper: use interactive map or infographic showing impact of issue
- Case study: use podcast or video with roleplaying showing real life decision making
- Art critique: use digital portfolio with images and audio recorded clips
- Consider multimodal assessments:
“AI will compel us to reconsider our approach to instruction and assessment, placing greater emphasis on process and authenticity, while prioritizing the depth of learning over the breadth of content. Consequently, we must exercise creativity in crafting our instructional methods and evaluations, while recognizing that this will vary depending on the discipline and course.” (Auburn Online, 2024).
- Give opportunities for growth
- Scaffold and provide support, encourage self-reflection and peer feedback, allow resubmission for growth (where appropriate)
- Asking for AI in assignments? Teach students professional use cases for ChatGPT. Encourage students to critically reflect on the material that is outputted from ChatGPT, and highlight the importance of understanding going beyond just generating information from ChatGPT.
- Focus on learning not production
- Assess the process, not just the product
- Reflect/journal the learning process
- Concept map the content and process
- Document key insights/learning in process portfolio
- Consider ungrading where the feedback and learning is center and not the grade
- Ideas:
- Use open book tests and move toward open ended conceptual learning questions, rather than memorization or review questions
- Have students submit a "my favourite mistakes I made" memo when using AI – mistakes and what they learned as part of their assignment
- Ask your students to provide an oral memo or complete an oral check in about their learning and their process as part of the assignment or for your feedback
- Communicate clearly, explicitly, and often
- At beginning: Syllabus Statement
- With each assessment remind students of guidelines
- Ask students about their experiences with AI
- Acknowledge presence of AI, explore it with students
“ Remember, the goal is not to create an environment where students feel constantly under surveillance or mistrusted, but to emphasize the value and rewards of their own work.” (Auburn Online, 2024).
- Value in-class active learning
- Rethink how you use class time (ex. time to do oral check in's)
- Use graphic organizers, drafts, brainstorming, group teaching
- Could an assessment be done in class?
- Involve students in live, graded discussions that require spontaneous responses that can't be outsourced to AI
- Value tools for specific purposes
- AI can help with:
- Collecting and organizing information
- Generating ideas, samples, examples, questions
- Providing quick feedback, explanations, definitions
- Supporting critical thinking, building arguments
- Engaging creativity, formatting documents/ presentations
- AI has limitations:
- Biases
- Inaccurate fact checking
- Causes invalidation of some types of assignments
- Causes questions of authenticity of student work
- AI can help with:
Now What?
Now it's time to pull all your ideas and thoughts together. It's a big task to revamp assessments. So, start small. Pick one assessment or assessment style this semester. Make the changes you are wanting. Get feedback from students.
As we all know, AI tools are not going away. AI tools will become even further integrated, customized and effective. We are already seeing AI supports built into Microsoft, Google, and BlackBoard products. Soon, multimodal AI tools will be combined into one with easy input and output (text, voice, picture, touch). We also truly hope there will be legislative and judicial points to address ethical concerns.
What does this mean for us? The only constant is change.
Want to learn more?
Consider exploring one of these resources:
- The Power of Authentic Assessment in the Age of AI - Faculty Focus | Higher Ed Teaching & Learning [new tab]
- Futurepedia - Find The Best AI Tools & Software [new tab]
- AI/ChatGPT Lessons - Using ChatGPT in the Classroom (Shared) - Google Docs [new tab] - Ideas for AI in Writing Assignments
- AI Misuse Checklist (Turn It In) turnitin.com/papers/academic-integrity-in-the-age-of-ai-misuse-checklist [new tab]
- Developing AI Pedagogical Practices | Tech & Learning (techlearning.com) [new tab]
-
Data Sets, Software & Tests - APA (7th ed.) Citing Guide - Library and Academic Success at Georgian College [new tab]
Resources
Auburn Online, (2024). Module 5: What Do I Need to Rethink in Terms of the Exams, Papers, or Projects I Assign?. In The Biggio Center for the Enhancement of Teaching and Learning at Auburn University, Teaching with Artificial Intelligence. catalog.auburn.edu