Reflections
The survey across campuses of the students showed a balance between useful application, testing, and confusion. A majority of the students said they used AI on specific tasks which they considered time-consuming, such as outlining, summarizing, citation, grammar, and clarity checking dense readings. Some students used AI to plan slide deck content and to develop simple visual concepts. A noticeable minority said they used AI to generate and submit large portions of their coursework, which they claimed required very little effort to modify. A significant number of respondents from this group raise apprehensions relating to, the accuracy and bias of the system, and the destination of the prompt and the data they provide to third party services.
These responses illustrate that what matters for ethics pertaining to AI in coursework is how the tool is applied. In cases where students use AI to organize their concepts, clarify dense passages, or surface edit their work, such with fact verification and authorship, the AI serves a learning purpose. In cases where the AI is a substitute for the students’ reasoning, who is unable to either justify the fabricated claims or the facts which are made to appear plausible, and the final polished work comes from the AI and the student is unable to explain the process, these are cases where the academic value is compromised and the student is unable to demonstrate proficiency. The survey responses showed that it was not just the use of a tool that was at issue, but rather a pervasive and inconsistent grasp of the concept of where assistance is welcomed, and where substitution is unwarranted.
To manage the uncertainty, I will implement focused and sensible approaches. I will leverage AI for the trim and robotic aspects but retain the analysis, interpretation, and logical construction. Before deep reading, I may request a short summary so I can orient myself, then proceed to deep reading and note-taking directly from primary sources. Once I draft a working thesis, I will ask AI for a couple of outline options for the thesis and then select one to rewrite and expand so the outline reflects my reasoning for the structure and how the transitions work. I will leverage AI for marking the work to flag grammar, clarity, sentence-level issues and all the major structural changes I will do myself. This way, I can maximize the usage of AI while being competent myself.
Within a time period of my routine, I will have the honor of being transparent. I will append a one-sentence statement to the submission, which says that if AI significantly alters the writing part, structuring, or idea creation, plus there’s some light editing, then I have to explain the kind of support received. This submission could say: “AI-powered outline; analysis and the final draft was synthesized by the student.” I will also maintain a minimal record for important tasks that contain the instructions, the results, and the changes made. This contributes to illustrating the thought process behind the ending piece as well as providing answers that could potentially be more complicated.
One’s image and bias are concrete risks that require relevant proficiencies. I would not insert proprietary survey answers, confidential student information, or private messages into commercial AI systems. I would AI-check any generated materials for bias, covert and overt, identity, demographics, and history and would ensure that disputable assertions are well supported by primary and secondary materials, relevant course readings, peer-reviewed articles, or other scholarly materials. For projects that require the use of sensitive information, I would use institutionally approved methods or submit requests to my fellow instructors and relevant stakeholders for permission.
