A Student Told Me They Were Ethically Opposed to AI. Here’s What I Wrote Back.

An instructor forwarded me an email from one of their students this term. The student wrote that they were morally and ethically opposed to AI. They were also worried about the environmental footprint of generative models, and about the communities that live near data centers. They asked whether the AI in Packback could be switched off, or whether they could submit their assignments some other way. In other words, can a student opt out of using AI in a required course?

It’s a question a lot of instructors are likely to hear over the next few semesters, and we can give students more insight than a simple reference to the course policy. It’s also not an isolated concern. So I want to share how I helped this instructor respond, and why we approached it that way.

Most of what students use in Packback is not generative AI

Students hear “AI” and picture the heavy hitters like ChatGPT, Claude, or Gemini. That’s understandable. For most people, those are the tools that have come to define AI. But the instant feedback, scoring, and moderation students encounter most often in Packback run on rule-based systems and predictive machine learning. That technology sits much closer to a spam filter or a spell checker than to a large language model. It runs on ordinary infrastructure and draws a small fraction of the energy a large generative model consumes.

We build it that way on purpose. For each job on the platform, we reach for the simplest, least energy-hungry tool that can do the job well. Predictive models are also faster, cheaper to run, and easier to explain to a student who wants to know why they got the feedback they got.

So when a student asks whether they can opt out of the AI, the honest first response is that most of what they are opting out of does not work the way they think it does.

What is the environmental footprint of AI used in education?

Let me concede the substance of the student’s concern before I complicate it.

Large generative models carry a real environmental cost. Training runs consume enormous amounts of electricity. Inference at scale adds more. Data centers draw water for cooling, and the communities hosting them absorb the local consequences of that draw and of the grid capacity those facilities claim. Reporting on this has grown more rigorous over the past two years, and the picture it paints does not flatter the industry. Students who are concerned about the environmental impact of AI have good reason to be.

That impact is a significant part of why at Packback, we lean on predictive systems wherever they will do the job. When we choose a lighter model to handle scoring and feedback, we are making an engineering decision and an environmental one in the same breath. 

Both things can be true: the environmental concerns around generative AI are real, and the footprint of the specific tool a student is using in their course may be much smaller than what they’re picturing.

To provide a sense of scale:

• A full year of Packback’s hosting emits less CO2 than the passengers on a single round-trip commercial flight between Denver and Chicago.
• Packback cut CO2 emissions from its operations by 34% between February 2023 and February 2025.

I share these figures to help put the choice in front of a student into perspective. The question becomes less about participating in the AI industry as a whole and more about the impact of this specific tool, in this specific course, and whether its use is worthwhile.

The generative features, and who controls them

A few features in Packback do use generative AI. The Instructional AI assistant is the clearest example. We label those features inside the platform so students always know which one they are using.

Faculty decide how much weight those features carry in your course. In the section that prompted this email, the instructor had already set up assignments where engaging with the generative features was optional, and students could earn full credit without touching them. That was a choice the instructor made when setting up the course, not a platform default, so check how your assignments are configured before responding to a student.

How should faculty respond when a student objects to AI on ethical grounds?

If a student raises a concern like this, start by taking the concern seriously.

1. Explain the distinction between predictive and generative AI. Students are not being unreasonable. They just have one word for two very different technologies, and a clear explanation may resolve most of their concerns. 
2. Check your own assignment setup before you answer. Look at whether the generative features are load-bearing in your course, or whether a student can complete the work without them. Answer from what you actually see.
3. Commit to flagging generative work ahead of time. If a later assignment asks students to use a generative feature, plan to proactively talk through how they want to approach it. Students accept a lot when they are not surprised.
4. Point to the documentation rather than paraphrasing it. Packback publishes an AI Ethics Policy and a plain-language breakdown of the technology behind each feature. Send the links and let the student read the primary source. Our team can also answer specific questions about data handling.

Why I am glad the student asked

A student who reads about data center emissions, connects it to the software their course requires, and writes a careful email to their instructor about it is doing exactly what we want students to do with technology. 

✅ They interrogated a tool instead of accepting it. 

✅ They asked who bears the cost. 

✅ They looked for an alternative before they complied.

That habit is the whole point of teaching students to work alongside AI rather than around it! Students who scrutinize a discussion platform will scrutinize a model’s output, a source it cited, and a claim it made confidently and wrongly. We should be building courses that reward that instinct, not courses that treat it as an obstacle.

So if a student asks you – answer the email. Answer it with real information. And then tell the student that the reflection and critical thinking behind their question are the skills the course is trying to build!