Designing Instruction & Assessments

When considering the design of instruction and assessments, there are a number of issues to consider. For now, this page will focus on:

Educator use of AI in instructional design and assessment
Student use of AI in completing assignments and demonstrating learning

Considerations for Educators:

Generative AI could help educators save time and energy by shifting some of their focus on lesson planning and assignment/project creation, to more attention placed on instruction, assessment, and determining the extent to which students demonstrate evidence of learning. Educational institutions are generally addressing two major topics:

Leveraging AI in planning instruction and assessment design

Mitigating the use of AI by students (plagiarism, misrepresenting learning, academic misconduct, etc...)

1. Leveraging AI:

Generative AI can help teachers with instructional suggestions and both general and specific brainstorming. With a few simple prompts, an array of ideas can be generated that would otherwise take an enormous investment of time and energy. 'Magic School' is a recently-introduced generative AI platform that offers teachers an array of tools for building rubrics, word problems, project ideas, unit plans and much more.

There's no question that generative AI is very powerful. It could help educators to...

suggest creative ways to introduce a new topic
outline steps of processes like labs, solutions, writing and more
demonstrate a concept or idea with unique content (write a paragraph on 'skateboarding' using a limited amount of ‘to be’ verbs)
generate individual inquiry questions for each student in a classroom

Whether using ChatGPT, Magic School or any other AI platform, be reminded that generative AI programs may not be accurate or correct in the responses they provide. Educators should always use their expertise and experience to determine the extent to which any AI suggestion might be implemented in the classroom. When determining whether the use of AI is appropriate, some issues educators will want to consider include cultural issues/context, community, social/emotional learning and the use of technology in the classroom.

In the fall of 2023, the University of British Columbia published a website entitled, 'Teaching and Learning in an Era of Generative AI'. This site reminds the user that the purpose of the learning should play a very important role in determining the use of AI:

'When considering revising assessments based on the availability of AI tools, it’s important to align the assessment with your learning objectives: what is the assessment activity measuring, and does this fit with the learning outcomes for the course? When deciding between various assessment options, consider whether students have achieved the learning goals of the course.'

Mitigating the Use of AI:

Throughout the history of education, there are times when students have misrepresented their learning by directly or indirectly claiming the ideas of someone else as their own. Whenever new technology is introduced, it could be argued that accessing the ideas of others gets easier. Some examples from the past include encyclopedias, Wikipedia, Google, and more. With many of these technologies, discovering that a student plagiarized an assignment involved finding the information that the student used. Generative AI is different. With the advent of systems like ChatGPT, it's very important to understand that the material generated is unique - it typically cannot be found anywhere.

When addressing how students might use AI in projects and assessments, and to mitigate the use of AI when it would be considered inappropriate, here are a few things to consider:

1. Assessments done in-class

In order to accurately determine the authenticity of student learning, educators may want to base their reporting on assignments and other representations of learning done in class.

The introduction of generative AI has caused educators to question some traditional learning activities and assignments. For instance, some educators are contemplating questions such as, 'Is the traditional take-home research paper dead?' and 'How can we assign writing pieces, math problems and other projects to be done outside the classroom when these demonstrations of learning can be easily fabricated?'

Longer assignments such as essays may not be feasible to do in one class period, but getting students to write a few paragraphs might be. Establishing the role of homework as practice, and then having students demonstrate understanding the next day in class may help send the message that using AI for homework is not a sound strategic decision on the part of the student. If we want to discourage the use of AI in completing homework assignments, it's likely wise that we not use the results from these assignments in determining a student's grade.

To be clear, there are many ways that students could misrepresent learning material produced elsewhere as their own. Generative AI is but one of these mechanisms. Therefore, as suggested by UBC, whenever possible student learning should be assessed in the classroom.

2. Human interaction followed by reflection

Students can bring something from their community, such as a photo or article, and incorporate it into their learning. Students can also collaborate with others and engage in a group conversation or activity, followed by an individual reflection (preferably done in class). When students engage in a group activity such as debates or simulations, generative AI programs do not have access that activity because 'they weren't in the room when it happened.'

Perhaps the introduction of generative AI could actually increase the number of learning activities that involve human interactions and face-to-face communication.

3. Personalize assessments

Assignments that incorporate student experiences greatly assist in mitigating the use of AI. When students are asked to relate an assignment to their experiences, family, or community, these unique and personal variables make using AI challenging. For instance, it's easy to use AI to summarize the main developments of World War 2. Alternatively, it's much more difficult if we ask a student to answer: 'How were young people in Penticton impacted by the second world war?' or 'Discuss a specific impact of World War 2 on your family'.

4. Non-textual demonstrations of learning

There are many ways for students to demonstrate learning. E-portfolios, videos, podcasts, photo collages, audio files, mind maps and infographics are examples of ways that students can demonstrate learning without using text. Currently the leading generative AI programs are text-based and the more we incorporate visuals, and particularly personal ones, the more we can mitigate the use of AI.

5. Experiential learning

The greater the variety of elements that contribute to a learning environment, the more it will be difficult for students to access AI for assignments and assessments. The following approaches/ideas make it difficult for students to rely on generative AI:

Outdoor learning
Incorporating a community guest speaker
Building/constructing
Conducting a simulation
Interviewing a local community member
Organizing a lab or experiment
Inquiry learning
Blogging
Hosting a mock trial or debate

With each of these approaches, students are required to apply knowledge to the activity while demonstrating skills such as communication, problem solving and creative thinking. Schools are entrusted to help students become knowledgeable, active citizens while functioning with the current technology of our era. When we focus on developing our students' competencies and skills it's possible that the acquisition of knowledge can come from the teacher, internet, AI or other sources.

6. Include the learning journey as part of the learning product.

It can be very tempting to base assessments on a final product such as an essay, set of answers or project. Another way to mitigate AI is to ask students to show elements of the process of their learning. If it's an essay, ask students to include rough drafts, an edit history and the research experience. If it's a hands-on project, ask students to develop an e-portfolio to document the learning journey.

7. Establish a baseline for student writing.

Generative AI platforms can quickly and easily produce pieces of writing based on simple prompts. If we continue to consider student writing to be an important skill, schools will need to take specific steps to ensure students truly engage in the writing process. To determine the next steps for any student's development as a writer, it's critical that educators determine the current skills and capacity of each student. In order to get students accustomed to being responsible for their own ideas and communication, here are a few suggestions:

Educators can give students the opportunity to do an in-class write early in the course/year, and frequently throughout the year. Once a baseline of writing is established for each student, it can then be compared to writing pieces practised or produced outside the classroom. If there is an obvious discrepancy, the educator can engage in a conversation with the student.
Establish continuous learning dialogues between the student and educator. These frequent conversations can help students set goals and communicate about their writing journey. Face-to-face communication between the teacher and the student can be an important step in mitigating the inappropriate use of AI.
Learning journals can be a very effective way to focus on the process of learning rather than just the product. Once students realize that learning involves the journey as much as the destination, we can seek to document the process through journals, e-portfolios, conversations and other ongoing forms of documenting learning. These personal progress reports are a very effective way to mitigate the use of AI.

8. Focus on the purpose of the BC Curricular Competency

Educators need to be very aware of the PURPOSE of the learning when determining the extent to which AI might be used. When examining the curricular competency, it's important to focus on the verb in the standard.

Generative AI is adept at coming up with accurate responses when given prompts that focus on lower-level verbs such as define, list, outline and identify. It's interesting to note that these verbs used to be the focus of educational curriculum.

Globally, the verbs in educational standards are changing. As learning activities shift in focus to higher-order verbs such as analyze, determine, model and construct, some argue that AI will struggle to produce quality responses. As time goes on, we may seek to infuse AI into our learning environments. AI might be of great assistance in obtaining the material that will lead to analysis, judgement, creativity and evaluation. For instance, ChatGPT can very quickly generate claims and counter claims. Once a document has been created from a prompt such as, 'What are 5 challenges faced by modern fruit farmers in Canada?', students can research the validity of the ChatGPT response.

This type of approach will require teachers to be very familiar with the learning standards and make the appropriate decisions as to when, why and how AI might be used to facilitate the process.