Artificial Intelligence (AI) is a game changer in the educational landscape. AI is a the latest technology that has the power to disruption education and specifically our current models of online education. Initially, we were looking at strategies to "beat AI", but after some research it can be more helpful to explore how to design courses that are AI resistant.
A side benefit of building AI resistant courses will be an improvement in the online courses overall. Designing assessments, learning objects and assignments/projects with AI resistance in mind will force course development down pathways that will enrich the current online course offerings.
The AI resistance philosophy to course building is not to suggest that AI use cannot be leveraged to be an educational benefit. In fact, AI may prove to be an incredible tool for education. The challenge for educators is in designing environments and experiences that promote greater learning opportunities and not simply allow students to let AI do their thinking for them.
Strategy Evaluation Table
AI Resistance Efficacy
(click on the strategy to jump to the explanation)
| Strategy | Efficacy | Pros | Cons |
|---|---|---|---|
| 1. Personalization of assignments/questions |
★★★★★ |
+ simple to implement + can increase student engagement |
- simple for AI to address |
| 2. Alternative Questions/Assignments | ★★★★★ |
+ simple to implement + can increase student ownership |
- ineffective against AI |
| 3. Evaluate/Make Decisions | ★★★★★ |
+ moderately simple to implement + can increase student engagement + higher on Blooms Taxonomy + AI struggles with decision making |
- creative use of AI can lead to evaluation |
| 4. Secure Browsers | ★★★★★ | + reduces access to outside resources |
- relatively easy to work around - tech skills needed |
| 5. Technological Supervision | ★★★★★ | + reduces access to outside resources |
- costs - third party provider |
| 6. Face to Face Supervision | ★★★★★ |
+ limits access to outside resources + AI Proof |
- resource intensive (staffing, location) - time consuming - reduce student control |
| 7. Video Conference Supervision | ★★★★★ | + limits access to outside resources |
- time consuming - students can "beat the supervision" |
| 8. Interactive Learning Objects | ★★★★★ |
+ limits the effect of outside resources + increases student engagement |
- costs and time to develop |
| 9. Teacher Presence |
★★★★★ |
+ increases student engagement + create clear expectations regarding AI |
- challenging in an asynchronous environment |
| 10. Outlines/Drafts | ★★★★★ | + provides students opportunities to plan and refine their thoughts |
- AI can easily produce outlines/drafts |
1: Personalization of Assignments
Including assignments that require personalization and reflection from the students can increase student engagement as students may feel greater relevance and responsibility. Tailoring assignments in a way that allows students to express their individual perspectives, experiences, and interests can encourage students to think critically about their learning experiences, connect new knowledge to existing understanding, and evaluate their progress.
AI Resistance Effectiveness (★★★★★)
AI can easily replicate personal responses. Through some simple, creative prompts AI can produce text that replicates reflection and can be tailored to include specific details that are relevant to the individual student.
AI itself will recommend the use of personal prompts to create assignments/questions that are less susceptible to AI. The issue is that with a few careful inputs into AI, AI can indeed replicate a personal response.
For Example: Using descriptive language, write a paragraph of a favourite memory you had while travelling.
As I sat on the bus at night, looking out the window, I felt lost in my thoughts. The passing streetlights made shadows on my face, and the soft sound of the engine was calming. Outside, the city lights sparkled like faraway stars, shining light on bits of the world around me. The quietness on the bus gave me time to let my mind wander, thinking about different things and daydreaming. I wondered about the people I saw through the windows and what their lives might be like. It felt like a moment to think and be alone, like a little break from reality. Looking out the window, the world seemed both far away and close, making me think about all the things that could happen in the future.
2: Alternate or Adjusting Questions?
Providing alternate assignments can increase student ownership by introducing more choice into the learning process.
AI Resistance Effectiveness (★★★★★)
Adjusting questions/assignments will be ineffective against AI. Unlike sites like Course Hero, AI does not provide plagiarized information, but provides original responses to the prompts.
Alternate assignments are also not effective in reducing AI use. AI can respond to unique prompts and so simply providing alternate choices is not a solution.
3: Evaluate/Make Decisions
Evaluation sits at the top of Bloom's taxonomy and is an opportunity for students to demonstrate several skills and understanding of the material while defending or explaining their decision.
AI Resistance Effectiveness (★★★★★)
Currently, AI's largest shortcoming is in evaluating or making decisions, but this also seems to be quickly changing. While initially when asked for a decision AI would provide a specific rote explanation of why it can't make a decision and then provide some general response. Currently, depending on the question, AI may indeed make a decision, or will provide a list of factors to consider in the decision.
4: Secure Browsers
Browsers that lock out student access to other tabs, or software can help the students focus on the task at hand and reduce the temptation to seek outside support.
AI Resistance Effectiveness (★★★★★)
While secure browsers can provide a layer of technological support to keep students focused on providing their own thoughts, a student could simply use their phone, or another computer to search out the information. The secure browser will reduce the copy and paste method, but students will still have relatively easy access to AI or other outside sources.
5: Technological Supervision
There are online proctoring services that will supervise students in an online environment. Through tracking eye movements, video monitoring. secure browsers these services can help ensure that the student is not utilizing outside sources.
AI Resistance Effectiveness (★★★★★)
Technological supervision can be very effective in ensuring that students are producing their own work. There is a cost for these services and districts would need to ensure they deal effectively around student data and privacy.
6: Face to Face Supervision
Face to face proctoring is the most effective solution in ensuring that students are not using outside sources.
AI Resistance Effectiveness (★★★★★)
Face to Face proctoring is resource intensive, but is the best method to insure that students are producing their own work and not accessing outside sources.
The effectiveness of this method does assume that the proctor is independent and providing quality supervision. (I am thinking of my own experiences when a TOC simply posted a password on the whiteboard for a test and told the students to complete the test when ready, or when a parent completed a test for their student.)
7: Video Conference Supervision
Supervising through Zoom, Big Blue Button, or other video conference solutions can be an effective method to ensure students are producing their own work. The flexibility of the online supervision allows students to complete the assessment at home at a time that works for teacher and student.
AI Resistance Effectiveness (★★★★★)
This solution is again resource intensive as supervision is frequently performed in a one to one environment and can be difficult to supervise multiple students at one time. Students can also design an environment that allows them the opportunity to try to subvert the supervision.
8: Interactive Learning Objects
Creating learning objects that students work through in the online environment can increase motivation and keep them engaged into the online environment.
AI Resistance Effectiveness (★★★★★)
Learning objects in themselves don't prevent students from using AI or outside sources, but provide the environment in which students are more likely to work through the objects on their own rather than take the extra time of jumping to AI and back to the learning object. If the learning objects are geared to more formative than summative assessment students will have little motivation to search out other sources.
9: Teacher Presence
Students who have a personal connection with their teacher are more likely to engage in the course, ask questions, communicate and provide their own work. Teachers setting clear expectations around plagiarism, AI use and assessment practices will help students understand the importance of producing their own work and the consequences if they do not.
AI Resistance Effectiveness (★★★★★)
Teacher presence can be more difficult to create in an online environment, especially asynchronously. The distance from the teacher and the student control of pace and progress through the course can lead students to feel that they are more "on their own" and not as invested in their education which could lead to more use of AI or outside resources.
10: Outlines/Drafts
Student outlines and drafts allow students the time and space to plan their thought process and explore and refine their arguments.
AI Resistance Effectiveness (★★★★★)
Outlines/Drafts are ineffective when it comes to AI use. AI excels in providing outline, and through some simple prompts it can produce drafts with mistakes.
There is no silver bullet in creating a course that is AI resistant. The combination of the above strategies will aid in the creation of a course that is resistant to the use of Artificial Intelligence and create an effective online environment.
If you are interested in exploring the use of AI in Education check out the "Positive Usage" page.