Strategies for Effective Feedback in the Age of Generative Artificial Intelligence

Authors: Patrick Kelly, Lorelei Anselmo, & Lin Yu
Updated October 8, 2024

Graphic of circuits around a person's head

Overview

This resource aims to help instructors at the University of Calgary navigate the use of generative artificial intelligence (GenAI) tools to provide feedback on student assessments. It discusses ethical considerations regarding GenAI’s applications in feedback processes and includes further tools and resources to support instructors.

Before using GenAI to generate student feedback consider the following

1

GenAI tools do not provide human judgement, but focus on structure and generic meaning (Banihashem et al., 2024)

2

GenAI tools may not be able to accurately process alternate media such as figures, images, tables and charts/graphs (Celik et al., 2022)

3

Intellectual property policies at UCalgary state not to upload student work to non- supported services without their permission.

4

GenAI tools do not incorporate UCalgary’s GenAI guiding principles found in these AI in teaching and learning resources

Since GenAI can exhibit biases, come across as impersonal, lack context to the assignment, and miss human judgement it is a good idea to stay away from using AI to evaluate and provide feedback on student work. Dr. Sarah Eaton explains these factors and more:

Further Reading

Why educators should avoid using AI apps to help with assessment

One day (soon!) we may have AI apps that can help with assessment of student learning but we are not there yet. For now, there are good reasons not to use AI apps to assist with the assessment of student learning. Here are a few.

Intellectual Property

A student’s work is their intellectual property. Unless you have permission to use it outside of class, then avoid doing so.

Privacy

A student’s personal data, including their name, ID number and pother details should never be uploaded to an external app without consent. 

Data security

Content uploaded to an AI tool may be added to its database and used to train the tool. 

Bias

AI apps are known to be biased. Feedback generated by an AI app can be biased and unfair.

Lack of Context

An AI app does not know your student like you do. It can provide generic feedback but may not help to scaffold a student’s learning. 

Impersonal

AI apps can provide generic feedback, but as an educator, you can personalize feedback to help the student grow. 

Academic Integrity

Educators model ethical behaviour, this includes transparent and fair assessment. 

Your Employee Responsibilities

If your job description includes assessing student work, you may be violating your employment contract if you offload assessment to an AI app. 

Adapted from Eaton, S. (2024). Read more at https://drsaraheaton.com/2024/09/10/ethical-reasons-to-avoid-using-ai-apps-for-student-assessment/

Supportive Tools

  1. Use UCalgary supported tools such as D2L or TopHat to automate feedback where possible. 
  2. Use Gradescope, an AI powered, UCalgary approved, assessment management tool. Gradescope can be used to help gather assessments, including hand written ones, match them to a class roster and then simplify the grading process, by grouping related assessments and constructing reusable rubric scores and comments. 
  3. Use AI to help design rubrics. Consider SMARTIE, a suite of AI powered web-based applications and Microsoft Copilot (sign in with your UCalgary IT account), an AI companion which can be prompted to generate rubrics.

Learn more about effective prompt writing to maximize AI at the link below. 

Learn more

What makes good feedback?

Effective feedback plays a crucial role in enhancing student learning and development. Nicol & McFarlene-Dick (2006) highlight 7 principles of effective feedback to promote student growth and learning.

  1. Effective feedback helps clarify what good performance is (goals, criteria, expected standards).
  2. Effective feedback facilitates the development of self-assessment (reflection) in learning.
  3. Effective feedback delivers high quality information to students about their learning.
  4. Effective feedback encourages teacher and peer dialogue around learning.
  5. Effective feedback encourages positive motivational beliefs and self-esteem.
  6. Effective feedback provides opportunities to close the gap between current and desired performance.
  7. Effective feedback provides information to teachers that can be used to help shape teaching. 

Strategies for effective feedback

With these principles in mind, here are some key strategies to ensure your feedback is clear and supportive.

Clarify expectations

Review the assignment with students, discussing the criteria and providing exemplars when possible. Clear expectations enhance transparency. For more on transparent assignment design, visit: Transparent Assignment Instructions Template.

Provide Class Time for Practice

Allow time during class for students to practice and receive additional feedback, such as peer or class-wide input. This helps students assess their abilities before submitting their assignments.

Scaffold Feedback

For higher-stakes assessments, have students submit a short proposal to receive quick feedback, guiding them toward the larger assessment.

Align Assessments with Learning Goals

Use constructive alignment to ensure all assessments accurately measure the course learning outcomes.

Consider Assessment Quantity and Weighting

Evaluate the number and length of assessments to reduce grading and feedback time, making the process more manageable.

Use Rubrics

Enhance consistency in grading and feedback by implementing rubrics. Explore the different kinds of rubrics and additional feedback strategies.

Final thoughts

Effective feedback is a crucial part of student improvement and student learning. While generative AI tools offer promising capabilities for enhancing feedback processes, it is crucial to approach their use with caution.  

Instructors should prioritize ethical considerations, adhere to institutional policies, and remain aware of the limitations of AI, such as its inability to provide nuanced human judgment and handle diverse media formats.  

By leveraging UCalgary-supported tools and integrating AI thoughtfully, educators can enhance the efficiency and effectiveness of their feedback while maintaining the integrity and personalization essential to student learning.  

Ultimately, combining AI tools with proven feedback strategies will help create a supportive and transparent learning environment that fosters student growth and development. 

References