Today’s notes are from “Teaching Responsibly with AI”a course offered by GPTZero and ran by Sean Michael Morris, Head of Pedagogy at GPTZero.
#1
How do you distinguish between authentic human writing and GenAI generated text?
Spectrum of Human-Synthetic (AI) Generated Writing
Human writing
Human-in-the-loop writing (HITL)
Machine-in-the-loop writing (MITL)
Synthetic writing
Before using AI Detectors (i.e. GPTZero), Consider:
How do AI detectors support your learning outcomes and goals?
What’s the purpose of using AI detectors, and can you articulate this purpose to your students?
How might AI detectors disrupt or enhance trust and transparency in your student-teacher relationships?
Avoid surveillance pedagogy. Use AI Detectors to start a conversation.
Notes:
- GenAI usage is invevitable in university nowadays.
- There's an ongoing debate on whether rules for using GenAI should be established in the context of semester assignments, theses and so forth.
- No official guidelines exist today on a university, national or international level to the best of my knowledge.
- "Avoid surveillance pedagogy. Use AI Detectors to start a conversation." -> This is a really good approach. Education should push students to make good use of the available tools, not penalizing their use.
#2
How do you set the tone of the learning environment in your classes?
Music playing; while people come to class.
What uses of GenAI cause the following:
Critical inquiry (reward to critical enquiry and course values).
Pain point (fails to meet our values and expectations, a minor inconvenience).
i.e. Citations used in assignments by students do not exist due to AI hallucination.
Harm (use violates our scholarly values, removes human from critical inquiry and problem solving).
i.e. When the paper is fully done by AI this is a harm since no experience was gained by the student.
Notes:
- "Music playing; while people come to class." It is useful! I am playing Baroque music during the lectures as well and it seems to have helped students focus better.
- The challenge is to lead students (reward?) to want to use GenAI to benefit from it rather than "get the job done" and move on.#3
How might academic integrity complaints without thoughtful conversations with students, about inappropriate generative AI use, be the antithesis of engaged pedagogy?
File complaints after a process of understanding
the context of inappropriate use,
the student’s reflection on how GenAI use did or did not further the values of the learning community and
GenAI’s risk to student learning for self-actualization and world transformation
How does the use of AI impact the learning of the student?
i.e. if the student over-relies on AI, they don’t learn from experience.
Rhetorical listening
understanding the experience of someone’s perspective of the world to arrive at truth
the antithesis of passive listening, which often seeks to further evidence and win a conversation from your own single perspective
emphasizes understanding, diplomacy, and communicating across cultures to seek ethical and justice-informed decisions
Notes:
- The concept of "Rhetorical Listening" is interesting!#4
Other approaches:
Metacognition
Before using GPT Zero: Ask students to annotate their draft for you, focusing on the rhetorical choices
After using GPT Zero: Students who meet a threshold of suspected GenAI use (for example, 6% probability), revise and annotate their draft for you, focusing on the rhetorical choices they made along the way
Reminder: The benefit of college education is doing deep internal work about your failures and successes, not just getting work done.
Notes:
- This seems to be a good first approach on solving the problem of not penalizing the use of GenAI but rather hold the student accountable for using it.
- By using thresholds the educator legitimizes their potential request to the student for the essay to be re-written.
- On the oter hand this tactic may lead the students to try and find better ways to "fool" the system, instead of doing the actual work.References: