Three Ways to Think About Ethical & Effective AI Use in Education by S. Tran

sntran57 — 2025-07-02 18:40:39 UTC

The AI Assessment Scale (AIAS; see site for research papers describing in greater detail) was introduced in 2023 and has been updated and used widely as a means to communicate the permitted levels of use of AI during an educational assessment. The purposes of the scale are to facilitate open dialogue between educators and students about appropriate use and support educators as they (re) design assessments in the AI era (Perkins, Roe, & Furze, 2024).

sntran57 — 2025-07-02 18:44:58 UTC

EDUCAUSE outlines eight ethical principles to guide the implementation of AI in higher education:

Beneficence: Ensuring that AI is used for the good of all students and faculty.
Justice: Promoting fairness in AI applications across all user groups.
Respect for Autonomy: Upholding the rights of individuals to make informed decisions regarding AI interactions.
Transparency and Explainability: Providing clear, understandable information about how AI systems operate.
Accountability and Responsibility: Holding institutions and developers accountable for the AI systems they deploy.
Privacy and Data Protection: Safeguarding personal information against unauthorized access and breaches.
Nondiscrimination and Fairness: Preventing biases in AI algorithms that could lead to discriminatory outcomes.
Assessment of Risks and Benefits: Weighing the potential impacts of AI technologies to balance benefits against risks.

These principles target specific ethical dimensions pertinent to AI use in higher education, emphasizing the importance of fair and equitable use, protection of individual privacy, transparency in decision-making processes, and a balanced assessment of risks and benefits. These principles must be viewed holistically and each leads to additional questions as institutions consider best practices for the responsible use of technology. For example, Beneficience leads to asking “How do we measure and monitor the benefits and risks of AI in an educational setting?” and “What safeguards are in place to prevent harm--such as algorithmic bias--from AI?” While concerns about justice lead to asking, “How does AI impact different groups, and how can we ensure fair treatment?” and “What knowledge, skills, or competencies are measured through AI, and are they aligned with course or institutional outcomes?” See the Appendix on this white paper for more questions.

(Read more at er.educause.edu and find the full white paper here.)

sntran57 — 2025-07-02 18:51:45 UTC

Read Ethan Mollick’s “15 Times to Use AI, and 5 Not to” for his thoughts of when AI is useful given its current abilities.

2025-08-11 18:21:35 UTC

I agree with the idea that AI should be used for the right kind of task. Brainstorming ideas can be helpful, especially for students who freeze up when they just see a prompt and a blank page. I’ve seen students get the ball rolling on large projects just by having a tutor suggest a topic or a starting point, and that can be super valuable. The reminder to check AI’s work also matches my own emphasis on critical thinking. Part of learning to write well is learning to question and revise. Sooo, here’s where my weariness sets in. More often than not, I see AI replacing the very practice my students need most: doing research, reading, shaping their own sentences, and struggling through the messiness of brainstorming and drafting. These “hard parts” are where I think skill actually develops and if AI is doing that heavy lifting the student might get a cleaner “grammatically correct/corporate sounding” paragraph, but they lose the proverbial muscle memory that comes from writing it themselves. IMHO, part of learning is it’s developing the stamina to work through confusion and frustration. I can align with the principles about checking AI’s accuracy, using it for low stakes things or busywork, and seeing it as a flexible tool. But my pedagogy/job depends on students engaging deeply with the process, not just turning in a final product

CypressCollegeESL — 2025-08-11 19:11:14 UTC

Tonya Cobb

So much to consider! These eight principles warrant consideration and integration into my course. Balance will be what I strive for... The full white paper provides great practical examples. I particularly connected with the one about using AI to (potentially) help grade--more equitably and to let students know when I will do so.

2025-08-11 20:14:30 UTC

I agree that generative AI is great at quickly summarizing large amounts of information, like summarizing assigned reading assignments for students in college. I also agree with times Mollick recommends we should not use generative AI, like when "learning new information" and when effort is at least part of the point of the assignment. These stood out most to me.

As an English professor, reading was and is an integral component of my learning, and I am concerned that generative AI will be too tempting for students to use to complete their assignments, trivializing their own learning and downplaying the role of effort.

2025-08-12 15:35:45 UTC

The section on when not to use AI spoke to me. I really like #1 and #4: Don't use AI when the effort is the point (the struggle is part of the process) and when you need to synthesize new ideas/information. Getting my students to understand this will be challenging, but it's definitely one of my goals for fall.

jungml13 — 2025-08-12 16:54:58 UTC

These principles are important to consider, especially beneficence and justice, which many of us can agree are essential. When we aim for beneficence and justice, access to AI tools should be distributed equitably and encouraged for all users. However, we must also be cautious—AI can be misused in ways that cause harm rather than help. In addition to making AI available to everyone, we need to provide clear guidelines to all and engage students in discussions about ethical use. Topics such as how AI can support critical thinking, rather than replace it, are key. Promoting AI as a tool to assist—rather than do the thinking for students—can help increase productivity while preserving the value of the writing process. Without this guidance, there's a risk that students may rely too heavily on AI, potentially hindering the development of their own critical thinking skills.

2025-08-12 21:10:01 UTC

Mollick’s reminder that effort is the heart of learning really resonates with me. My hope is that my students will one day be the CEOs of their own lives, equipped with the knowledge, confidence, and critical thinking to guide their decisions and recognize the difference between good and bad work. AI is powerful, but it works best when we already have the expertise to see its strengths and weaknesses. I want my students to use AI the way a CEO delegates tasks, strategically and with oversight, so they can focus on leading. That means putting in the work now to become experts first, so they can lead with confidence and use AI wisely as a tool for success.

2025-08-12 22:42:26 UTC

This is very education-centric, which isn't always the best thing. There are concerns here that educators have no control over. The only thing we can do is question things like bias in AI; very few of us are in a position to actively do anything about it. I realize I may be quibbling about semantics here, but it bothers me that we teach students how to identify bias but not necessarily what they should do about bias.

afterschooluniv — 2025-08-13 02:06:17 UTC

Annette Letcher

AI has created a Catch-22 in academia; as stated in the EDUCAUSE article, "AI accelerates creativity but risks embedding bias, obscuring accountability, and eroding critical educational relationships." We certainly don’t want to stifle the creativity of our students, as AI can be a conduit for generating ideas for assignments and problem-based solutions. However, just as it can act as a channel it could also function as a drain on students critical thinking, natural curiosity and creativity. It is important to consider the eight principles from the EDUCAUSE ethical framework when introducing assignments and assessments in order to keep authenticity, responsibility, accountability, and equality flowing through the pipeline.

2025-08-13 05:25:24 UTC

(JF) The varied approaches proposed by the three leaders for ethically and effectively using AI provide valuable insights for implementing AI in the classroom. Although each writer provides their suggestions and rationales from a different perspective, a common ground of principles can be found across the three perspectives, thereby providing a baseline of best practices and principles for using artificial intelligence.

qiand1 — 2025-08-13 05:55:18 UTC

The model offers a valuable framework for thinking about how to integrate AI into teaching at various levels, moving beyond the binary choice of “all AI” versus “no AI at all.” One of the key challenges that accompanies this model is assessment, particularly at the higher levels of integration, such as AI collaboration, full AI, and AI exploration. These levels provide ample opportunities for creative uses of AI, but they also demand a reimagination of how to evaluate such substantial and innovative AI-based approaches in teaching. I think classroom-based research becomes more important than ever. It can provide stakeholders with a deeper understanding of what it truly means to embed AI into existing teaching practices, and guide the development of meaningful, fair, and effective assessment strategies.

qiand1 — 2025-08-13 06:39:42 UTC

I like this piece and plan to share it with my students. The second point, "work where you are an expert and can assess quickly whether AI is good or bad", is especially relevant to my teaching context. One challenge for many of my multilingual students is that they are not yet confident in seeing themselves as experts in English academic writing. They need targeted support to develop the skills necessary to effectively evaluate AI-generated writing outputs.

2025-08-13 12:04:27 UTC

One of the "don'ts" stood out to me, When the effort is the point. In many areas, people need to struggle with a topic to succeed - writers rewrite the same page, academics revisit a theory many times. By shortcutting that struggle, no matter how frustrating, you may lose the ability to reach the vital “aha” moment." I stressed during the meeting yesterday that showing one's thinking and work BEFORE you engage with AI is key to getting the most out of AI. I believe there needs to be a framework for learning when using AI.

Re-centering cognitive, epistemic, and ontological agency so students own their thinking and decision-making.
Affirming cultural ways of knowing so AI isn’t the default authority over lived knowledge.
Embedding ways to operationalize disciplinary skills as a meaningful part of mastery, so “aha” moments emerge from interaction, not automation. Plus, you build in check's and balances that are user driven.

pinventado — 2025-08-13 13:54:30 UTC

I really like the scale and how it can clarify your expectations to students. Many students hesitate to use gen AI for fear of getting labeled as cheating, while others use it without realizing how its use takes away from their learning experience. Additionally, I think it would be good to emphasize how you can use AIAS' different scales at different parts of a task. You may start out with No AI, then later switch to AI collaboration. This might happen inside a single class activity, or spanning across different activities.

pinventado — 2025-08-13 13:56:53 UTC

I think Ethan Mollick's framework is pretty useful, but I find it tricky to use, especially if you're new to it. There is a lot to consider, so I wonder if it can be synthesized in some way to help users. Perhaps a decision tree of some sort or other visual guides will help.

pinventado — 2025-08-13 14:03:59 UTC

I see myself using these principles as a checklist to help evaluate decision-making on AI. However, it seems like there may be some trade-offs when you try accommodating all these principles. For example, promoting Autonomy might affect Justice and Beneficence negatively if the user is not mindful of their AI decision-making. I wonder if adding notes about trade-offs among the different principles can help people use them more effectively.

2025-08-13 14:12:59 UTC

This is Roselyn Du (not sure if I need to sign up for a Padlet account to show my name for the post). Mollick's article makes an excellent future reference for me going into the new school year. I will certainly share it with my students, and I am sure they will find it informative and educational.

One thing that I would like to extend reading further from this article is "o1, the new AI model from OpenAI, can solve some PhD-level problems." I wanted to know what is defined as PhD-level problems and see some examples.

cassandramatsuya — 2025-08-13 15:07:05 UTC

I appreciate Ethan Mollick’s approach to AI, recognizing its usefulness while still being aware of its limitations. The point made about recognizing the value of expertise is something that I wholeheartedly agree with—the more refined the input, the more effective the output will be, which is why the critical thinking skills we pass onto our students in our courses are essential. However, I am still skeptical about relying on AI for summaries of written texts. Although Mollick acknowledges the potential for misinformation as one of his caveats, I believe close reading and full engagement with the original source material is a crucial aspect of learning. Will annotated bibliographies and academic research fundamentally change in response to AI? Will students still be able to effectively analyze and evaluate sources? What are we sacrificing for the sake of efficiency? How will research be limited or potentially enhanced by this technology? I recognize there may be room for a middle ground—having AI summarize to understand a larger concept and then doing a deep dive of select texts, but these are some questions I’m grappling with. Perhaps, this may align with Mollick’s point about “people need[ing] to struggle with a topic to succeed - writers rewrite the same page, academics revisit a theory many times. By shortcutting that struggle, no matter how frustrating, you may lose the ability to reach the vital ‘aha’ moment.” Ultimately, this aligns with my stance on critically engaging with peer reviewed sources; the challenge is part of the learning experience, and it is important for students to participate in this discourse with fellow, human authors.

tgideons — 2025-08-13 15:51:31 UTC

This felt like a good companion to the EDUCAUSE framework, because it integrated both ethics and pragmatism. As someone said in the discussion, it's a bit repetitive and leads to "but ... " responses. But I think it was helpful to think with, and I think it's something more easily shareable for working with students and mentees.

stakeuchi1 — 2025-08-13 15:55:14 UTC

This chart is interesting, and I think it provides useful language to help students understand the degrees of AI use in completing assignments. This will help assuage some of the anxiety they may feel about being asked to use but not misuse AI. I'm becoming more aware about how worried many students are, so having a simple chart that defines use will be a good teaching tool.

stakeuchi1 — 2025-08-13 15:58:30 UTC

These are important points to consider, but I agree with another colleague's point about what we then DO about some of these concerns. For example, transparency and explainability calls for the companies that create and run LLMs to share information that they probably won't in order to protect trade secrets, and this is further complicated by the black box problem--even the developers can't always explain how AI's decision-making process works. Though the article does provide more concrete examples of how each can be implemented on an institutional level, some still seem beyond one professor's ability to control.

stakeuchi1 — 2025-08-13 16:00:06 UTC

I appreciate this list for its potential to be adapted for students and used as a way to clarify what is acceptable or unacceptable use of AI. A colleague of mine at Chaffey College created a decision-making tree to help their students navigate ethical use, and I think this list can be used to create a similar decision-making tree for my own students.

2025-08-13 21:05:00 UTC

I appreciate this frame work, especially the ways in which not to us AI. Mollick mentions that "asking for a summary is not the same as reading for yourself." What I have noticed, with students to rely on AI in the classroom is that some have this thought process. They are students, after all, and the purpose of being in school is to learn. Part of their learning is to understand when we need to strengthen their own writing skills and when AI is helpful in that process.