AI-Aware Implementation: Assessment Examples

Incorporating GenAI into this course allows students to experience realistic use of GenAI in a translation context. Comparing their own translation with that generated by a GenAI tool helps emphasize evaluation and analysis skills (higher-order cognitive tasks).

PART 1:

• Translate an expository text or literary excerpt in the language of your choice.
• Write a reflection that explores your translatory choices and how they might reflect the “accuracy” vs “overall meaning” translation theories.

PART 2:

• Compare your translation to that of Google Translate’s or another GenAI model.
• Write a reflection on the translatory decisions that GenAI makes, including which translatory decisions you might keep and why.

Modified from Professor Per Urlaub’s example presented here.

Using GenAI to assist with complex programming enables students to spend more effort on conceptual understanding of the problem. This allows the instructor to focus assignments on more complex and realistic chemical engineering scenarios and analyses. It also makes the course more inclusive since proficiency in MATLAB programming becomes less important.

Modified from Professor Michael S. Strano’s example presented here 

Case Study: Helping the Boston Metro School District erase a $50 million deficit

The case method, which asks students to analyze existing case reports and evaluate proposed recommendations, is a cornerstone of business school learning. While valuable, this approach does not fully develop the skills consultants need in practice.

In this assignment, instructors use an AI-based simulation of a realistic consulting scenario (e.g., helping the Boston Metro School District address a $50 million deficit), which allows students to interact extensively with multiple stakeholders, each with distinct roles and perspectives, and gather data, analyze findings, and develop evidence-based recommendations.

AI Is Teaching the Next Generation of M.B.A.s the Classic Case Study & Noyes AI example

Writing is broken down into stages:

Detailed assignment here

Writing assignments are scaffolded and broken down into multiple smaller assignments that include student reflection and peer participation in class. Breaking down large summative assessments into smaller components with support and guidance makes it easier for students to complete the assignment, increases their motivation by boosting self-efficacy and agency through feedback and iteration, and reduces the likelihood that students will cognitively offload their work to AI.

At the end of each unit:

• In small groups, students create (1-class session):
  • List of terms & concepts
  • Questions using these concepts

• Students chose one piece of their own writing (couple of class sessions):
  • Analyze & respond to Q’s
  • Brainstorm answers
  • Peer review

Detail description of approach here

Students are told that although AI can be used for certain tasks, students should not offload their learning to AI; “Collaborate! Don’t Delegate!” is the instructor’s mantra to convey this message.

Students are provided with case studies representing various stages of the research and writing process. In small groups, they decide whether the case studies are examples of collaboration or delegation. Together, the instructor and students decide on a comprehensive AI policy for all the class’s learning outcomes.

This approach allows for a transparent conversation about learning outcomes and provides students with agency, making adherence and motivation to follow the policies more likely.

Example of AI-use case studies

For each task, permission of AI usage is designated into one of 3 categories based on student collective discussions:

Laura Roberts’ AI Tool Box for Research & Writing

This assignment was designed to illustrate the limitations of GenAI use for certain tasks. Students are asked to reproduce a donut plot in R using two different approaches: GenAI assistance and instructor’ provided guidance. GenAI tools struggle to reproduce donut plots accurately, demonstrating that sometimes it is better to accomplish a task without relying on GenAI.

1. Reproduce a donut plot in R (statistical computing & graphical language)
2. Compare both attempts.

Detailed assignment here