Adapting Assignments for the Age of AI

Generative AI tools (genAI) are platforms that use natural language prompts to generate new, human-like output. Students are using AI tools when they feel pressure to get good grades, lack confidence, or want to save time. In addition to navigating valid concerns about bias, privacy, equity, and academic integrity, instructors are adapting their assignments and teaching methods to a new AI reality.

The widespread availability of genAI tools poses challenges for cognitive engagement, academic integrity, and assessment design. One study found that people who used genAI for writing tasks showed less cognitive engagement than those who wrote independently, igniting concerns of diminished critical thinking skills. While these findings don’t make AI inherently harmful, they highlight how easy it is to offload cognitive work to this technology. Because these tools can quickly generate passable essays and code, many traditional out-of-class assignments no longer accurately measure student learning.

Fortunately, there are strategies to adapt assessments for the realities of an AI-integrated world.

When adapting your assignments, begin by reviewing your course goals and learning objectives to ensure they remain appropriate for the AI era using these reflective questions. You can also learn more about creating learning objectives in CATE’s Learning Objectives teaching guide.

Beyond assignment redesign, you can support students by offering clear guidance, setting expectations, and modeling transparency through your AI policy. You can build trust by sharing how you use genAI tools in your teaching and research. Use CATE’s syllabus guide to create a transparent syllabus.

AI Detection Tools

No assignment is completely “AI-proof,” because detecting AI use is fraught with technical and ethical challenges. Researchers have previously found AI detection tools to be inaccurate and biased, yet these tools are rapidly evolving, with some platforms demonstrating greater accuracy. Given these mixed results, UIC does not support AI detection tools.

Instructors should only use AI detection tools to understand patterns in student work, not as definitive indicators of AI misuse. Instructors should also notify students if they plan to use AI detectors, as they can easily create distrustful learning environments. Thankfully, there are many strategies that, when combined, can reduce the likelihood that students will turn to AI as a shortcut. The most productive teacher-student relationships are transparent and collaborative.

Technology Integration

UIC-supported technology platforms, including genAI tools, are vetted for ADA compliance and FERPA data privacy requirements. When using third-party genAI tools (including AI detectors), be cautious about uploading intellectual property, copyrighted material, and student data.

Regardless of how you integrate gen AI tools in your course, we encourage you to spend time building both general and discipline-specific AI literacy skills with your students. While there are many ways to accomplish this, a simple first step is to work through appropriate examples together, followed by a discussion of inappropriate use cases. Always verify AI-generated output before sharing it with students.

We will describe three common assignment types found in college courses: 1) foundational knowledge recall, 2) skill application, and 3) cumulative essays and projects. After each type of assignment, we will provide strategies you can use to preserve the benefits of these assignments despite the challenges posed by genAI, as well as strategies that leverage AI tools to enhance student learning.

Each strategy is flexible and can be adapted across multiple disciplines and assignment types. We encourage you to find a balance of strategies that preserve academic integrity and facilitate technology integration. You might find that scaffolding or layering multiple strategies can support learning and engagement while building critical AI literacy.

Assignments that prompt students to recall facts and define terms are crucial for any discipline.

Example formats: Multiple-choice, multiple-answer, fill-in-the-blank, true/false, short-answer questions

Example uses: These assignments can serve as formative assessments given throughout the term to help both instructors and learners understand how well they are doing, or as summative assessments to measure mastery.

Example learning objectives: The following learning objectives focus on recalling foundational knowledge from across multiple disciplines:

• Biology: Describe the different muscle fiber types in the human body.
• Marketing: Identify the elements of marketing research.
• Sociology: Define the concept of “cultural capital.”
• Graphic Design: List the primary, secondary, and tertiary colors.

What is the value of this type of assignment?

Internalizing foundational knowledge helps learners build a mental library they can use to master complex, higher-order tasks.

Students need to practice recalling core concepts so they aren’t just looking things up, but truly “know their stuff” as confident, independent thinkers.

How to preserve the benefits of this type of assignment?

Prioritize low-stakes quizzing: Short, low-stakes quizzes can reduce test anxiety and provide learners with additional opportunities for feedback. When these quizzes are incorporated at consistent intervals (e.g., after each reading or before discussion sections), recall practice becomes a purposeful habit that can reduce the likelihood that students will turn to genAI.

Incorporate carefully placed knowledge checks: To encourage practice outside of class, consider using Canvas quizzes to create short, frequent assessments that can be assigned directly after a reading and automatically graded (possibly with multiple retake options) or within a video using Canvas Studio or Panopto. When it is clear that the purpose of the knowledge check questions is to reinforce mastery before moving on to more difficult concepts, students are less motivated to shortcut their learning with tools like genAI.

Assignments that require students to apply knowledge in context are essential for building disciplinary fluency.

Example formats: Problem sets, case studies, coding assignments, annotated bibliographies, reports

Example uses: These assignments can serve as formative checkpoints, where students receive feedback and instructors gain insight into how well students are learning, or as part of a summative assessment to measure mastery.

Example learning objectives: The following learning objectives focus on applying skills and completing tasks from across multiple disciplines.

• Nursing: Calculate the correct dosage of a medication based on a patient’s weight and age.
• Psychology: Analyze a given longitudinal study of human development for potential confounding variables.
• Statistics: Compute descriptive statistics for quantitative variables, including central tendency (mean, median, mode) and variability (range, standard deviation).
• Music Theory: Transpose a given harmonic progression to a new key.

What is the value of this type of assignment?

Assignments that ask students to apply skills are valuable because they help learners move beyond knowing what something is to knowing how to use it.

It is the “productive friction,” or cognitive struggle, that requires students to apply their knowledge while developing the critical thinking and problem-solving skills they need to be successful in any field.

How to preserve the benefits of this type of assignment?

One of the best ways to preserve the benefits of these types of assignments is to ensure learners are engaged in relevant, real-world tasks. These types of authentic assessments improve student motivation and make it less likely they will turn to genAI as a shortcut. Similar to recall assignments, frequent, short, low-stakes assignments will also provide learners with additional ways to practice targeted skills.

Pre-assignment planning: After sharing a general description and overall goals, but before sharing the specific details of an assignment, ask students to identify useful resources, describe potential issues, or draft a plan of action. This “human-first” step allows instructors a window into how students plan to apply their skills and prompts metacognitive thinking, as students begin to reflect on what they already know. If the task is completed during an on-campus course meeting or as part of a small-group discussion, students may be less likely to outsource it to AI.

Incorporate oral assessments: Instructors can require that students verbally describe the process they used to complete an assignment, either in an oral exam or informal interview. Whether or not students use genAI for a task, verbalizing a process requires a clear understanding, strong mental models, and internalized logic. For smaller classes, this may include one-on-one discussions with TAs at several key points throughout a course. For larger classes, instructors may require students to sign up for one or two short interviews throughout the term using a video conferencing platform.

Externalize thinking: Ask students to complete a task or problem set, prompting them to show their work. After they complete the task, require students to annotate their process. When built into an assignment, these annotations encourage students to closely examine the process or concepts they are studying, and to complete the task as intended without supplementary AI tools. The following is one possible prompt to add to a practice assignment.

• Sample Instructions: Annotate Your Thinking: After completing the problem, add at least three annotations to your work. Each annotation should address one of the following prompts:
  • Decision: Explain why you chose this method or approach.
  • Course Resource: Provide a link or citation to the course resource that you referenced when completing this step or section.
  • Troubleshooting: Identify a moment where you initially made a mistake, explain how you diagnosed it, and how you corrected it.

Require reflection: Instructors can also supplement application assignments with brief post-assignment reflective writing tasks that ask students to describe what they learned, what was challenging, and what they still have questions about. Keeping these reflections short and focused on personal insights makes them more meaningful for the student and less likely to be outsourced to AI.

Assignments that require students to produce long-form writing or multi-component projects are essential for synthesizing knowledge and skills surrounding a specific topic.

Example formats: Essays, research papers, large-scale designs, capstone projects

Example uses: These assignments are typically used as summative assessments at the end of a term to measure how well students have synthesized knowledge and skills into a comprehensive artifact or presentation.

Example learning objectives: The following learning objectives focus on synthesizing ideas or creating coherent new structures from across multiple disciplines:

• Computer Science: Develop a complex software application that addresses a specific user need.
• Theater: Critique a modern adaptation of a Shakespearean play based on the way it balances authenticity with contemporary relevance.
• Public Health: Develop a public health campaign to address a specific health concern for a local community.
• Sociology: Evaluate the role of media in the construction of social identity among adolescents.

What is the value of this type of assignment?

The real value of essays and projects lies in the process. In many disciplines, the final project is where learners begin feeling less like a “student” and more like a “practitioner.”

Synthesizing knowledge and integrating complex skills enables students to bridge gaps and build more durable, flexible mental frameworks.

How to preserve the benefits of this type of assignment?

One of the best ways to preserve the benefits of cumulative essays and projects is to ensure learners are engaged in relevant, real-world tasks. When students understand how an assignment clearly aligns with their future careers, they are more likely to complete it as intended. The following strategies can help preserve the benefits of cumulative essays and projects and can be adapted for other assignment types as well.

Write in class: “Blue books,” or in-class handwritten exams, can maintain assessment integrity and often assess unpolished writing more than take-home assignments, which allow more time for thinking and editing. When considering this format, instructors should weigh factors such as handwriting skills, varied thinking speeds, and neurodivergent learning profiles. To manage these challenges, some instructors are experimenting with in-class writing on laptops with minimized distractions (e.g., no Wi-Fi).

Scaffold submissions: Instructors can break up large assignments into smaller, sequential pieces. For example, a final research proposal might include several submissions spaced a few weeks apart to allow for meaningful peer or instructor feedback (e.g., a rationale, a literature review, a method, etc.). This approach shifts the focus from the final product to the learning process, fostering critical thinking at every stage and reducing the likelihood that AI will produce cohesive deliverables.

Highlight growth: Instructors can require students to compile their work throughout a course into a portfolio. As a collection of artifacts, this assessment can show growth by illuminating a student’s evolving knowledge and skill set over time, something that genAI tools are unlikely to mimic convincingly.

A final summary and reflective statement also allow students to highlight their growth while connecting individual assignments into a cohesive whole.

Embed metacognition: When students think about their own thinking, they build problem-solving and self-regulation skills. The personal nature of tasks like this can motivate students to value their own learning and complete assignments as intended. For assignments that already exist in parts, consider adding multiple layers of metacognition, including multiple moments to reflect on struggles, successes, and growth at different stages.

Require process tracking: Prompts students to share their drafting process for a writing or large-scale project. You can offer students options for how to do this, including a collaborative document with saved versions (e.g., Google Docs), a Grammarly Authorship report, or prompt students to think of another way to highlight their process. These strategies promote transparency about AI use during the drafting process.

Assign alternative deliverables: Instructors can break a large project into pieces spanning multiple formats (e.g., text, slides, video, live presentation). While AI tools have multimodal capabilities, they don’t always maintain consistent arguments and logic across platforms. These inconsistencies require students to translate their knowledge into different formats, mirroring dynamic professional environments where multimodal communication is essential.