What Students Should Know About AI: One Professor’s Thoughts About Ethical AI Use

Disclaimer: no text or image has been generated, revised, or altered by AI.

What’s the narrative?

The first thing to know is that AI has excellent PR. Take this commercial from Anthropic, for example. It’s a slick rhetorical move to persuade the viewer that questions like Is AI safe? and Can AI be trusted? and Do we even need this? are the same kinds of questions as Could AI help me feel included? and Can AI care for me? This commercial from Meta claims that AI has always been about connection and humanity. The tone, rhythm, images—it’s practically inspiring. If you haven’t seen those commercials yet, maybe you’ve seen the ads for Amazon’s Alexa featuring Pete Davidson or Chris Hemsworth or Lil Wayne? AI is fun and funny!

But these commercials have me wondering about a few things: (1) Why does an AI company or AI tool need PR like this? Is there something they want to pull my attention from? (2) What even is AI? The commercials make AI look like a slightly flirty machine that will fix my sink or call an Uber for me, but it also seems like something that could provide therapy or even companionship. Or maybe it’s just job tools? Is it Google? Is Google AI now?

Stop saying “AI.” Name the action, instead.

A report commissioned by the United Nations acknowledges that defining AI is a “moving target” that has “shifted over time, from symbolic AI to machine learning, to generative AI, agentic AI and sometimes even artificial general intelligence or superintelligence.” It’s difficult to have a conversation about AI or to express caution or reticence about the benefits of AI when the definitions are so murky. In fact, being skeptical of AI often feels like standing against progress like some 21st century Luddite (a group who, historically, had really bad PR). For example, the UN’s report shares that an AI called AlphaFold “predicted the structures of more than 200 million proteins, now used by over 3 million researchers” to accelerate “drug design, vaccine development and antibiotic resistance research.” Surely, we shouldn’t be against this kind of AI, right? (And, I’m not!)

Dr. Avriel Epps, a specialist in AI ethics and justice, suggests that instead of saying “AI,” we should call AI by the thing that it does. Instead of “AI,” say “cancer modeling.” Instead of “AI,” say “text/image/music generation.” Instead of “AI,” say “facial recognition.” And so on. The biggest PR win for AI is that instead of identifying specific AI platforms or tools and the particular things those AI agents do, we call everything “AI” which lumps the scary, the questionable, and the good all together. If we want to be thoughtful and productive students, teachers, and educators when it comes to AI, we need to be more specific about the actions and tools we do or do not want our students to use. AI isn’t a monolith, and it’s helpful to talk through what we do and don’t like about it.

AI is biased.

It’s just a machine, right? Right? Someone built the machine, though. Even asking the machine to learn isn’t neutral because the large language models (LLMs) and source texts have a point of view and positionality. The UN report mentioned above found that AI development is highly concentrated. A few countries have immense influence on the direction, development, and use of AI. This has at least two major implications for AI outputs: a set of unbalanced values; and, a hegemonic disposition. 

Rather than exploring a range of values and interests, AI is currently focused on business development, with marginalizing effects for most people—surveillance, data tracking, but also the nature of the tools, themselves, are not so much about creativity and innovation as they are about efficiency and (micro)management. (There are also a ton of social and environmental issues like data centers, water use, and information security, but that’s content for a different essay.) Even worse, AI is not culturally neutral. AI advantages already-dominant socio-economic groups and can be actively racially biased. When we ask AI a question or to complete a task or to write something for us, the result is generally from a wealthy, business-positive, English-language, and white perspective. AI has not been developed from an intentionally inclusive environment. Students, teachers, and administrators using AI to generate text, music, or images, to do research, or to find sources may find that the results aren’t representative of the cultures, languages, and lived experiences that exist in our classrooms.

The bill comes due as cognitive debt.

Early studies of AI use within education and the classroom (such as these from MIT, Carnegie Mellon University and Microsoft, and the University of Graz in Austria) suggest using AI leads to “cognitive debt.” Certain tasks, like writing or doing research, don’t benefit from using AI in the same way that other subjects or tasks do. Compare this to students using a calculator (a technology that offloads the cognitive task of arithmetic) in a math class. Generally, though, before students are allowed to use calculators, they are required to first learn how to manually accomplish the calculation. But beyond that, the calculator enables the student to engage with more complex mathematical concepts. Essentially, the calculator allows the students to do more difficult work. Instead of less thinking, students do deeper thinking. For writing and research (and similar tasks), the offloaded task is the thinking. That is, writing and research are not simply labor that needs to get done to arrive at some other more complex task. Writing and research are the work of thinking.

When students use AI to write an essay, to find sources, to summarize, or even to brainstorm, they are offloading the primary educational task (by the way, these are our student learning objectives…). Even more concerning (than missing an educational opportunity), the research suggests that students actively lose the ability to complete the original task. They get worse at writing, research, summary, and other similar tasks, perhaps irrevocably so. And, going back to the math example, if students use AI or a strong calculator to simply solve problems instead of to facilitate calculations or check work, the cognitive debt shows up, too, because the cognitive task has been outsourced and offloaded. Additionally, the research reveals that students using AI report less personal satisfaction with learning. Doing the work of learning doesn’t always feel great, but having learned feels good, and it seems that too much AI in the learning process steals that joy.

Be a critical thinker.

The only way to respond to potential cognitive debt, institutional bias, and a strong PR machine is to be a deliberate and methodical critical thinker. More thinking is the antidote to the tools that allow us to avoid thinking! First, don’t offload any tasks to AI that involve thinking, judgment, or decision making. Leave those for your own mind. When you do use AI, though, be slow about it. Question the results. Do they make sense? AI has been known to spit out nonsensical sources and answers that are simply wrong. Also ask: where are the AI responses coming from? Who benefits from the content? What ideas are present and which ones might be missing? What do I gain or miss out on from not doing this task myself? Slow, thoughtful, deliberate. That’s the key to productive and ethical AI use.

Just because AI is here, doesn’t mean it can’t be better.

Sometimes people will say things like, “AI isn’t going anywhere, so we might as well hop on board.” That’s not a compelling argument. It ignores the fact that most people are more concerned than excited about the increase of AI. It’s also defeatist. Dr. Epps reminds us that “we have to challenge what we don’t want so that we get the future we do want” when it comes to AI use and policy.

This essay may seem anti-AI, but it’s not. I’m not anti-AI. Instead, I want us to be careful and conscientious users of technology. I also want more people voicing their ideas and concerns, hopes and goals when it comes to the future of AI and the future we are trying to build as learners and global citizens. TL;DR: What should students know about AI before they come to college? They need to know that using AI is a complex issue implicating everything from copyright issues and plagiarism to environmental activism to our own capacity to think, act, be creative, and to participate in the important conversations that determine the paths of our lives and the lives of those around us. It’s not that deep. But it is.