A common pattern is seen in online discussions: students paste a discussion prompt into tools like ChatGPT or Gemini, submit the generated answer as their own post, then use AI again to reply to their classmates. The result? A conversation of bots talking to bots: efficient, maybe, but not conducive to learning. 

The good news is that you can design discussion boards that invite authentic student thinking and make AI shortcuts easier to spot. The five strategies below focus on ways to alter your discussions in a way that prioritizes student thinking, application, and interaction. 

Note: Not every strategy will fit every course or class size, so consider piloting one or two approaches that align with your learning goals and workload.

Use multimedia to change the medium of expression

Discussion boards don’t have to be all text. Asking for submissions and peer responses in other forms of media adds friction, makes AI easier to spot, and makes authentic participation the easier path. 

Something to try: 

• Narrated two-minute videos (with captions) explaining a concept, experience, or reflection. Peers can respond with a video reflection of their own. 
• Annotated images or infographics that visually represents an argument, process, or comparison. Peers can provide embedded feedback by notating directly on a copy of the submission. 
• Brief podcast-style reflections sharing insights or relevant stories. Peers can reply by submitting a “caller question” to that post. 

Keep in mind: Multimedia assignments can create accessibility and technology barriers for some students. Be clear with your expectations and allow alternative options to help ensure equity. 

Ask for highly specific, personalized, contextual responses 

Discussion boards can be tailored to unique contexts. Consistent individualized detail is difficult for AI to fabricate, especially over the arc of an entire course. Students are also more likely to engage actively in discussions that directly apply to their own lives, interests, and/or professional aspirations. 

Something to try: 

• Ask students about something in their local community and have them describe the issue in detail, proposing one actionable improvement. 
• Have working students analyze data from their own department and propose a solution. 
• Ask for reflections about personal experience related to the topic area, what made it memorable, and how it shaped their approach to learning or professional growth. 

Keep in mind: Even personalized prompts can be gamed with creative prompting or fabricated details. Gather student context early (e.g., interests, workplaces, local issues). By getting to know your students and asking them to work through personal/local applications, you should be able to identify whether discussion posts accurately reflect their experiences and perspectives. 

Lean into authentic assessments

Lean into authentic assessments that ask students to apply course knowledge and skills to complex situations that mirror the real world. Because the work is tied to genuine contexts and the student’s own reasoning, AI often struggles with the nuanced choices, lived details, and process explanations required. Have your students tackle realistic, context-rich tasks and require “how” and “why” explanations. 

Something to try: 

• Designing a training plan for their workplace. 
• Analyzing a local issue. 
• Developing a community outreach proposal. 

Keep in mind: Authentic assessments often take more time to design and grade because they emphasize open-ended, individualized responses. Each student’s work looks different, which makes feedback more time intensive. Clear rubrics and exemplars can help ease the grading process and prevent overload. 

Require course-specific evidence

Ask students to support their ideas with specific references to course materials, such as timestamps in lectures, quotes from assigned readings, or references to previously covered topics. Because LLMs cannot access or recall your specific course content, students must engage directly with materials to respond accurately. Make sure students know these references are required components of participation. 

Something to try: 

• Referring to at least one example, quote, or concept from the week’s video lecture. Include the timestamp and explain how it connects to their post. 
• Citing a specific quote from one of the assigned readings that supports or challenges their argument. 
• Referencing a key concept or application from the previous week and explaining how it connects to the discussion. 

Keep in mind: Students can still attempt to mimic engagement by paraphrasing course concepts vaguely, using false citations, and/or superficially inserting course-specific references into AI generated outputs. You may need to spot-check citations and model what genuine integration of course materials looks like. 

Have students keep an AI usage log

Empower responsible use of AI. Explicitly allow students to use AI but require disclosure about how they use it. This transparency creates accountability, discourages wholesale copy-paste, and provides insight into the student’s learning process. Mismatches between the declared process and the final post can quickly flag questionable work. 

Something to try: 

• Notating the exact AI prompts they used. 
• Providing the raw AI output. 
• Describing how they revised or built upon it. 

Keep in mind: Some students may be hesitant to give full details of their AI use, especially if they fear penalties. This strategy works best when you foster psychological safety and frame the log as a tool for reflection, not surveillance. 

Conclusion

The goal is not to eliminate AI from discussion boards but to design in a way that encourages authentic participation. When you make a strong case that your discussion boards are worth their time and individual ideas, it becomes harder to automate. 

Ultimately, students are more likely to engage honestly if the discussion is dynamic, relevant, and uniquely theirs. When discussions spark curiosity and ownership, authentic participation outshines anything a bot can produce. 

---

References

• Bielousva, G. (2023, April 24). Creating AI-proof assignments: A guide for university professors in social sciences and humanities. LinkedIn. https://www.linkedin.com/pulse/creating-ai-proof-assignments-guide-university-social-gra%C5%BEina/ 
• Carleton College. (Summer 2025). AI-resistant assignments. https://www.carleton.edu/writing/resources-for-faculty/working-with-ai/ai-resistant-assignments/ 
• Colorado State University. Academic integrity: How do I AI proof my assignments? https://tilt.colostate.edu/how-do-i-ai-proof-my-assignments/ 
• Florida Gulf Coast University. (2024, July 2). GenAI proof your discussion forums. https://www.fgcu.edu/digitallearning/digital-learning-blog/2024-02-07-aiproofdiscussions