7 Strategies for Redesigning Assessment in Response to Artificial Intelligence

Higher education is changing fast, and generative AI like ChatGPT and DALL-E are shaking things up. These tools can whip up text and images from simple prompts, making you wonder if student work is truly their own and if traditional assessments still cut it.

To tackle this, you might need to rethink your assessment designs to minimize academic misconduct and boost learning. Here are seven key strategies that can help.

Suggested by the Melbourne Centre for the Study of Higher Education, these strategies aim to create more genuine and engaging assessments despite technological advancements.

But keep in mind that these strategies might not fit every situation perfectly. You’ll need to consider things like subject matter, student level, and class size. Plus, redesigning assessments isn’t always a walk in the park workload-wise.

Also, as you explore the strategies with practical tips, case studies, and resources; remember that the goal is not just to prevent misconduct but to enhance student learning in our ever-evolving educational landscape.

Strategy 1: Emphasize Process Over Product in Assessment

In your quest for more effective and secure assessments, shifting your focus from the final product to the learning process can offer significant advantages. This approach values not just the end result—be it an exam, report, or essay—but also the developmental journey your students undertake.

Process-oriented assessment evaluates the steps and strategies your students use during learning, giving you insight into their thought processes, problem-solving approaches, and self-reflection abilities. This method offers two key benefits:

Enhanced Insight into Student Learning: You gain a clearer understanding of how your students think and approach tasks, allowing you to provide more targeted instruction and support.
Development of Metacognitive Skills: Your students learn to monitor and manage their own thinking and learning strategies, fostering independent learning.

Moreover, assessing the process rather than just the final product can help mitigate concerns about academic integrity, as it’s more challenging for students to outsource their learning journey.

Practical Implementation:

Learning Journals: Incorporate self-reflection tasks (written or video) that prompt your students to evaluate their progress in relation to the subject’s intended learning outcomes (ILOs). Ask them to identify areas of confidence and challenges and to develop plans for improvement.
Process Notebooks: In practical or practice-based subjects, require your students to document their steps, actions, and learnings. This allows you to assess the inquiry, experimentation, or application process, rather than focusing solely on the final outcome.

Strategy 2: Incorporate evaluative judgment tasks

Assignments where you need to show your evaluative judgment by reviewing or assessing work—whether it’s published articles, peer-created content, images, objects, audio, or video—help you develop higher-order skills like knowledge application, evaluation, and critical thinking. While it’s not impossible to outsource these tasks, they are much harder to complete using generative AI.

Practical Implementation: Reflection through peer review

Peer review tasks get you involved in evaluating your classmates’ work, fostering reflection on quality and drawing from various inputs. You’ll provide constructive feedback to your peers and reflect on the feedback you receive for your own work. This might include demonstrating how you’ve applied the feedback to improve a draft or explaining how you’d use the knowledge in future assessments. There are various school-supported educational technologies (LMS and other software) that can help manage the peer review process.

Strategy 3: Design nested or staged assessments

This strategy focuses on the process by creating assessments that build upon each other throughout the semester, ending with a complex piece of work that shows you’ve nailed the subject’s learning goals. You’ll break a big assignment into 3-4 steps.

This strategy has a few benefits:

You get feedback (automated, peer, or instructor) after each step.
You understand how to plan and complete complex tasks better.
You get evaluated on how well you can apply knowledge and adapt plans based on feedback.
It’s tougher to cheat using generative AI because it includes group work and reflections on individual contributions.

Examples of Semester-Long Assessment Design:

Case Study Analysis and Recommendations
- Task 1 (early semester): Case Analysis - Identify key issues and apply relevant concepts to discuss a complex scenario.
- Task 2 (mid-semester): Recommendations - Develop evidence-based solutions based on your case analysis and feedback.
- Final Task (end of semester): Implementation Plan - Present a detailed plan for addressing case issues, which could be an oral presentation, poster, or written assignment.
Project-Based Group Work
- Task 1 (early semester): Individual Research - Find and summarize five authoritative sources on a problem your group identifies.
- Task 2 (mid-semester): Group Project Plan - Synthesize individual work, outline team roles, and justify your approach.
- Final Task Part 1 (end of semester): Group Presentation - Present your analysis and recommendations using a format like a poster or PechaKucha.
- Final Task Part 2 (end of semester): Individual Reflection - Critically reflect on peer and personal contributions and what you learned from feedback.

Strategy 4: Diversify assessment formats

Incorporating a variety of assessment types can help you reduce vulnerability to academic misconduct via generative AI while offering your students diverse opportunities to showcase their learning. Consider these approaches:

Video-based assessments: PechaKucha presentations or recorded poster sessions encourage visual thinking and oral communication skills.
Audio assignments: Podcasts or simulated expert interviews allow students to demonstrate research and communication abilities in a different medium.
Reflective video logs (vlogs): Ideal for documenting experiences in practical subjects or work placements.
ePortfolios: A collection of multimedia artifacts demonstrating achievement of learning outcomes.

These formats resist AI misuse and foster creativity and authentic skill development. However, be mindful of potential resource inequalities among your students when implementing non-textual assessments.

Strategy 5: Design authentic, context-specific assignments

Let’s make assignments that mirror real-world tasks or tie closely to your course material! This approach not only ramps up student engagement but also makes it tougher for AI to assist in cheating. Here’s how you can do it:

Analyze case studies or scenarios that directly relate to your course materials.
Evaluate lesser-known local features or objects relevant to your discipline.
Encourage personal reflections that draw on individual or peer experiences.

These approaches make assignments more relevant to your students while keeping academic integrity in check.

Strategy 6: Leverage in-class and group assignments

In-class assessments give you a versatile way to evaluate your students, using formats like quizzes, live polls, tests, concept maps, short written tasks, and oral presentations. You can design these for individuals or groups.

By using in-class tasks, especially collaborative ones, you create an environment that encourages peer interaction and mutual learning. Team-based activities during class time also cut down on opportunities and motivation for academic dishonesty.

Practical Implementation:

Peer and Self-Assessment of Group Work: Develop clear criteria focusing on group work processes like effective communication, goal-setting, active participation, quality of contributions, and respect for diverse views. Discuss these criteria with your students so they understand how to evaluate themselves and their peers.
In-Class Concept Maps or 5-Minute Papers: Use brief, low-stakes assessments throughout the semester to help students consolidate their understanding. Have them create concept maps (individually or in pairs) or write short papers explaining key concepts, applying knowledge to problem-solving, or reflecting on recent learning experiences.

Strategy 7: Implement oral interviews to evaluate understanding and knowledge application

Oral interviews are a powerful way to assess student knowledge while cutting down on cheating. By asking students to respond verbally to unpredictable prompts, you get a real sense of their comprehension and reasoning skills. Sure, it might be a bit more stressful for some, but it simulates real-world scenarios and helps students develop those crucial communication skills.

When dealing with large classes, you’ll need to plan carefully. Moderation meetings with multiple assessors can help ensure that everything scales smoothly and stays consistent.

Here’s how you can apply this:

Scenario- or Case-Based Interviews: Give students a brief scenario and ask them to identify key issues, explain the relevance to the subject, or answer specific questions. For complex cases, you can let them prepare beforehand.
Practice-Based or Procedural Interviews: In subjects that involve practical skills, have students explain step-by-step procedures for tasks, including relevant safety protocols.
Paired Interviews or Role Play: Pair students up and let them take turns as interviewer and interviewee on related topics. This is especially effective for subjects that need advanced oral communication and interpersonal skills.

Bottom line: As we all adapt to this new educational paradigm, see it as an opportunity to innovate and improve your teaching practices. Embrace the changes, continuously refine your assessment methods. This way, you can do your part in making sure that education remains meaningful and transformative in the age of artificial intelligence.