How AI is Changing the Script in Education

Note: This post was first written in conjunction with McGraw Hill Education

As one of the pioneers of the Flipped Learning movement and a leader in the Mastery/Competency-Based educational model, I have increasingly focused on how artificial intelligence can enhance the learning experience for students. In my exploration of AI, I am particularly excited about its potential as a student tutor. Additionally, AI assists me in creating quality assessments in my Mastery Classroom. In this article, I will share examples of how AI can streamline teachers’ preparation time and improve students’ mastery of content.

AI Helps Create Summative Assessment Questions

AI is assisting me with many of the administrative tasks that consume a significant amount of my time. I have extensively used AI to draft assessment questions for one of my classes. Since I teach using a mastery learning approach, students are required to retake an assessment if they do not master the material on their first attempt. However, this raises the question: how do you maintain so many tests in your filing cabinet?

For the past six years, I have been utilizing our Learning Management System to randomize the questions presented. Essentially, I create question banks and instruct the system to select a specific number of questions from each bank randomly. This way, every student taking the same test receives different questions that assess the same topic.

For example, in the teacher’s view of one of my tests, I have multiple banks of questions to draw from. In the “Earthquakes Short Essay” section, for example, the program randomly selects two questions from my bank of twenty. As a result, when a student takes the test a second time, they will likely encounter different questions.

One of the significant challenges of Mastery Learning is creating large question banks. Since I am teaching a new class (Geology, Meteorology, Oceanography — GMO), my question banks are not very extensive. Therefore, I decided to use an AI tool to help generate questions for me. In one of my prompts to ChatGPT, I asked it to write 30 multiple-choice questions based on the attached document, which was a PDF of a section from my textbook. This approach allowed the program to provide relevant questions.

After that, I followed up with another prompt: Can you now write 20 free response questions that increase in complexity based on Bloom’s Taxonomy? The AI generated some valuable questions to assess student knowledge. While not all the questions were suitable, I utilized my understanding of Geology to keep, modify, or discard questions as needed.

Although I use AI to enhance my assessments in a Mastery Learning classroom, this approach also benefits various teaching methods. Many educators at my school utilize the randomized test feature of our LMS, allowing students who retest to receive a different exam that still evaluates the same objectives. You might wonder about test security. Our school employs a lockdown browser that restricts students’ access to everything except the test, which helps maintain academic integrity.

AI Tutors for Differentiation and Help

In a previous article (linked here), I discussed how AI accelerates Mastery Learning, specifically focusing on AI tutors. Since that publication, many teachers at my school have begun using AI tutors with great success overall. As mentioned earlier, our school utilizes FlintK12.com to serve as an AI tutor. Teachers create activities and specify how they want the tutor to interact with the students. Here are a few examples from our school.

High School English Example

One English teacher prompted the AI to have students compare parallel stories. In her instructions, she first asked the AI to inquire about which book the student read (either “1984” or “The Picture of Dorian Gray”). The tutor then engaged the student in a dialogue covering four specific topics. The prompt was quite detailed, as you can see below. It’s important to note the level of engagement the teacher expected, and the AI successfully facilitated a rich conversation with each student.

High School Chemistry Example — Using AI in a Flipped Class

One of our teachers is teaching Chemistry via the Flipped Classroom model. Prior to AI, he assigned videos using EdPuzzle. One cool feature of EdPuzzle is the ability to have it monitor student video behavior. It tells you how long they watch a video, and then it allows you to ask multiple-choice questions to check for understanding. However, with the AI tutor (FlintK12.com), this teacher linked a video that he had hosted on YouTube and then had the AI ask a few open-ended questions. Not only does the AI grade their level of comprehension, but it globally looks at all students who watched this video and tells you the strengths and weaknesses of their grasp of the content. See below one of his prompts.

He can even drill down to each student and find out what their strengths and weaknesses are. Also, when each student completes a video, they can see their strengths and weaknesses. This level of both personalization and data should change the way we teach.

Limitations

AI is not perfect. It didn’t write perfect questions. When I asked it to make a large bank of questions, I discovered that it started repeating questions. It didn’t create unique enough questions for me. Also, it isn’t a perfect tutor. Some of our students have gotten misled or gone down rabbit holes and had very interesting conversations with the AI tutor. In one instance, a student started asking about black holes, and she learned lots of cool things.

In another instance, the AI graded one of this teacher’s top students, and she asked if she could answer more questions to receive a higher grade. What followed was a series of interactions where the AI seemed to get snarky and almost insult the reading level of the student. It started talking about something called the Gunning FOG index.

When I do presentations about AI to teachers, I tell them that AI tutors are like having a college student in my classroom who is there to help my students. The “AI college student” isn’t perfect it — can be a bit wordy, and it sometimes can go down rabbit holes. But…I now have as many tutors in my room as I have students. I just need to supervise the tutor and hopefully help the tutors help my students.

We truly are in a new era of education. There is so much potential with AI, and specifically AI tutors, but we have a long way to go. I am excited to see what other creative ways you might be using AI in your classroom. What have you found?