MrKremerScience.com

Research-supported instructional practices are teaching approaches that have been proven to be effective by rigorous and objective analysis of data from the learning environment (LINCS 2017). These practices can help educators design and deliver instruction that is aligned with the learning goals, needs, and preferences of their students. However, implementing these practices can be challenging, especially in large and diverse classrooms, where teachers may not have enough time, resources, or support to provide personalized learning experiences for every student.

This is where artificial intelligence (AI) can play a crucial role. 

AI is everywhere these days. AI, of course, is the branch of computer science that aims to create machines and systems that can perform tasks normally requiring human intelligence. AI has the potential to transform education by enhancing research-supported instructional practices in several different ways. Here are some of the innovative ways that AI can be used to improve teaching and learning:

Produce smart content

Teachers can use AI as a virtual assistant to help them create and curate high-quality and engaging content for their students, such as interactive simulations, games, videos, podcasts, and ebooks. With a few well-crafted prompts, AI can also source content that is tailored to the students’ level of understanding, interests, and learning preferences. For example, a teacher can curate relevant resources by inputting the subject, year level, content standard, assessment criteria or desired skills, then tasking the AI to suggest activities and resources at low, medium and high levels of mastery. The AI acts like a pre-filter for Google searches.

Create exemplars

In my IB Environmental Systems and Societies (IB ESS) class, students must write structured essays for their Paper 2 examinations. The essays are assessed according to a fairly complex 9-point rubric. Students must understand the rubric to be able to produce a high-quality Paper 2 essay, so I feed one of the AI platforms essay prompts from past exams. Once the essay has been written, the students and I use the rubric to assess the AI’s work. This accomplishes a few things:

• Students develop a deeper understanding of the expectations established in the rubric.
• Having the exemplar allows students to evaluate the quality of the writing. So far, they have enjoyed poking holes in the AI’s arguments and identifying parts of the rubric that the AI failed to address.
• They can see how formulaic most AI writing is and how surprisingly incomplete or incorrect it often is. This dispels them of the notion that computers and the internet are inherently reliable and may discourage them from using AI as a substitute for their own work in the future.

Differentiate tasks and activities

One of my staff was struggling to reach all the students in his language acquisition class. Because of the size of our school, the class by necessity had students of radically different ability levels in it: some were brand new, having never been previously exposed to the language. Others had been studying it for only one year. And a third group of students were almost fluent in the language, having studied it for 4 years or more in previous schools. How did AI help?

1. Our first prompt was to input the unit demographics: number of lessons, assessment criteria, inquiry questions, content standards and desired skill outcomes from the achievement descriptors in the written curriculum.
2. We then instructed the AI to suggest summative assessment tasks aligned with the assessment criteria and the language levels of the students. We then adjusted these personally to ensure they made sense and aligned with our local context.
3. Next, we fed each edited task back to the AI and instructed it to use McTighe and Wiggins’ Understanding By Design protocol to produce an outline of lesson topics that would build students’ skills for the summative assessment.
4. Then we told the AI to suggest two or three activities per lesson, stipulating that each activity must be based on one of Robert Marzano’s Instructional Strategies That Work or Barak Rosenshine’s Principles of Instruction. It was necessary to do this in batches of fewer than 5 lessons at a time so that we didn’t max out the AI’s output capacity.
5. Once we had a collection of activities for each lesson, we tasked the AI to locate free or open-source resources to support each activity.

This whole process took about 2 hours of trial-and-error with the AI – that’s all!

After all was said and done, my teacher had a 10-week unit plan with a detailed sequence of of lessons and differentiated activities targeting the students in his class. Whenever it produced something that wasn’t quite right, we simply adjusted our prompt and instructed it to re-write its’ previous output. By applying a station-rotation model to most lessons and occasionally pairing higher-ability students with new language learners, the teacher saw student learning accelerate notably.

There’s a lot more that AI can do to make teachers’ lives easier and their instruction more effective. These are just a few of the innovative ways I’ve found to enhance research-supported instructional practices using AI. AI still poses some challenges and risks that need to be addressed carefully. These include ethical issues such as privacy, security, bias, fairness, transparency, accountability; pedagogical issues such as validity, reliability, quality; social issues such as equity, inclusion, diversity; and human issues such as agency, autonomy, and empowerment.

My next exploration with AI is to dive into the world of AI-assisted Universal design for Learning (UDL).

And for what it’s worth, I used Bing Chat to do these things. Perplexity, Bard, or other AIs might produce quite different results.

References

Beale, Jonathan. “Barak Rosenshine’s ‘Principles of Instruction.’” CIRL, Eton College, 7 Jan. 2020, cirl.etoncollege.com/barak-rosenshines-principles-of-instruction/.

“Evidence-Based, Student-Centered Instructional Practices.” LINCS, The Literacy Information and Communication System at the U.S. Department of Education, 10 May 2017.

McDaniel, Rhett. “Understanding by Design.” Vanderbilt University, Vanderbilt University, 10 June 1970, cft.vanderbilt.edu/guides-sub-pages/understanding-by-design/.

Staff, TeachThought. “Marzano’s 9 Instructional Strategies for Teaching and Learning.” TeachThought, 16 Jan. 2022, www.teachthought.com/learning/instructional-strategies/.

“Generative Artificial Intelligence in Education: What Are the Opportunities and Challenges?” UNESCO.Org, UNESCO, 6 Sept. 2023, www.unesco.org/en/articles/generative-artificial-intelligence-education-what-are-opportunities-and-challenges.