By streamlining the evaluation process of student work, teachers can reduce the time spent creating targeted and customized feedback. Additionally, incorporating experiences that teach students to reflect on and assess their own work and that of their peers can be invaluable. These experiences develop critical thinking and self-assessment skills and foster a collaborative learning environment where students learn from one another's strengths and areas for growth. This approach not only lightens the teachers' grading load but also empowers students to take an active role in their learning journey. While AI support in the evaluation process can be beneficial, precautions must be taken not to lose the connection between student and teacher. 

WHY Is This Important 

There are several ways to use Generative AI (GenAI) to evaluate student work and create assessments. By leveraging the capabilities of large language models (LLMs), teachers can streamline the creation of assessments, ensuring they align with learning objectives and are accessible to students with diverse learning styles.

What Will Be Learned?  

  1. Reducing bias in grading 

  2. Prompting for student reflection 

  3. Teacher feedback and reflection

  4. Protecting student-data privacy  

Examples of tools that may be useful for this task:

  • ChatGPT Plus: This service requires a paid subscription. If you have access, the Explore GPT option provides access to a library of chatbots on various topics. The free version offers limited access to this feature. ChatGPT Desktop is available and functions similarly.

  • Flubaroo

  • Microsoft Forms

  • Socrative


Reflective Practices

References

Guiding Reflection Questions

Suggested Action Steps

1.

AI’s Impact On Teaching Practices - Automated Grading, Let Summarize Ed, 2023. (Video 3 minutes 23 seconds)

Why Should I Watch This?

AI and Machine Learning can support teachers' ability to grade and provide feedback to students automatically. Understanding how this automation works allows teachers to implement the benefits effectively.


How have automated grading and feedback tools impacted your teaching practices? If you haven't used automated grading and feedback tools, what benefits from the video would you see benefiting you most?

Ethical Use 

It is important to maintain student-teacher relationships around feedback. How would you maintain the relationships if AI generates the feedback? 

Conversations are happening around the benefits and risks of AI in support of teaching and learning. Read educators' thoughts on the automated evaluation and feedback potential of GenAI on Reddit  HERE.

LLM can be used to grade written assignments, requiring detailed rubrics and other prompting to achieve desired results. However, LLMs can also support automated grading for straightforward tests like multiple-choice. 

Practice using AI to grade a multiple-choice test automatically. 

Take the following steps within the LLM of your choice. You will need to have the following items ready to train your LLM. 

  • A quiz, answer key, and student responses.

  • Detailed grading criteria (points, percentages, total score. etc.) 

Upload your information into your LLM. 

  • Once you upload your information, ask the LLM to grade students' responses and generate feedback for incorrect answers. 

The Design and Automate Section contains example prompts for this process. 

2.

Feedback From an AI-Driven Tool Improves Teaching, Stanford-led research finds.  Government Technology, March 2023 (Article) 

Why Should I Read This? 

Much emphasis has been put on teachers using GEN AI to evaluate student work. However, teachers also need feedback on their performance. The administration often only provides feedback a couple of times a year. The article provides insight into the benefit of using GEN AI to support teachers receiving timely evaluation feedback on their instruction. 

Teachers are Using AI to Grade Essays, But Some Experts Are Raising Ethical Concerns CNN, April 2024 ( Article)

Why Should I Read This? 

The article provides insight into the integration of AI with tools like ChatGPT. It examines the potential benefits for students to critique their work and teachers to provide feedback and free up time for personalized instruction. It also examines current ethical and other potential concerns. 

How can you see GEN AI providing insights that could positively impact your teaching practice?

Ethical Use

When utilizing GEN AI to evaluate student written work, be careful not to upload identifiable and personal information about students. Also, talk with the administration about obtaining student and parent consent. 

Let’s use GEN AI to provide feedback for your growth and reflection by comparing our assessment with one generated by a chatbot you train. 

Choose a written assessment, the accompanying rubric, and a sample student submission for which you have already evaluated and provided feedback.

Upload the documentation into your chatbot and create prompts to train your AI Chatbot to provide a grade and feedback for the student's work. 

Mark Students Assessments, by Bron Earer (13 minutes 02 seconds), provides a walk-through on accomplishing this. 

Reflection: 

  • How did the AI's grading and feedback compare to your own evaluation?

  • What insights did you gain about the AI's potential and limitations in grading written assessments?

3.

AI and Assessment: Where Are We Now, AACBS April 2024 (Article) 

Why Should I Read This?

It explores ways others envision and use Gen AI to benefit the classroom. It explores ideas such as GenAI, which allows instructors to be more involved in the student learning journey and not just evaluate final products or grade summative tests.

How can you utilize Generative AI tools to free time to engage more deeply in students' ongoing learning processes beyond traditional grading and assessment methods?

Ethical Use

Teachers must remember that AI-generated evaluation has limitations and cannot provide the following, as detailed in Tom Daccord's book AI Tools Uses: A Practical Guide For Teachers. July, 2024

  • AI looks for patterns but does not comprehend meaning and nuance. An articulate but factually incorrect response could receive a good score.

  • AI can reflect biases present in training data.

  • AI  is still not able to assess highly creative original thoughts that deviate from expected response patterns, 

  • Teachers must have insight into the scoring results to discuss them with parents and students. 

AI can also make things up, which is referred to as hallucinations. The article below provides a detailed example of a hallucination's appearance in practice. 

Creative Applications of Artificial Intelligence in Education, Ed Tech, July 2024

Utilizing AI-automated grading tools that integrate with tools such as Google Forms and Sheets allows teachers to easily set up automated grading. It provides an option to train your own Chatbot in LLM, such as ChatGPT and Co-Pilot.

Let’s practice using the free Flubaroo tool with Google Sheets. The site's videos will walk you through setting up your automated grading.


Design and Automate

Creating student-friendly, meaningful language at students' reading level is part of crafting individualized feedback and guidance when assessing student work and writing. How can GEN AI be used to simplify and target the language we use for individualization based on specific student needs?

Once you have trained the Chatbot with the assignment, rubric, and other information to prepare it to assess student work, use it to elicit written feedback based on specific student-differentiated needs. Use the following prompt examples to support you. Here is an example transcript using ChatGPT. Free and paid versions of generative AI chat boxes may require different levels of prompting to complete the work. 

Prompt Examples:

1. Reading Level Identification:

  • How would you provide feedback to a 5th-grade student with a basic reading level?

  • What adjustments would you make in your feedback language for a student reading below grade level?

2. Specific Student Needs:

  • How would you provide feedback to an English Language Learner (ELL) to ensure they understand your comments?

  • What considerations should be made when providing feedback to a student with dyslexia?

3. Encouraging and Supportive Tone:

  • How can you phrase feedback to be constructive and encouraging for a student who struggles with writing?

  • How would you acknowledge a student's effort while suggesting improvement areas?

4. Differentiating Feedback:

  • How can feedback be modified to cater to students at different academic levels?"

  • Provide an example of feedback tailored to a student with advanced writing skills compared to one just beginning.

Additional sample prompts in the article 50 Time - -Saving ChatGpt Prompts For Teachers, Shana Ramin (Blog) 


Logo of Michigan Virtual, an organization that provides online learning resources. This logo is used to indicate external resources provided by Michigan Virtual. Example Prompt 1.6 Timely and Targeted Feedback provides a prompt on how AI can be used to deliver feedback efficiently, enabling educators to quickly identify student needs and provide personalized support to enhance learning outcomes. Prompt Example 2.5 Tutoring and Support explores the design of a AI math tutor to help students understand and master algebra concepts. Prompt Example 4.0 Administrative Task demonstrates the development of a set of email templates that can be used to create custom feedback.