Using It, Scared of It
Before we get into prompts and tools and outputs and all of the shiny things these platforms can do, I wanted the students in my GenAI course to pause and write honestly about what they already think and feel. What comes to mind when they hear “generative AI”? Where have they seen it? What have they used it for? What do they trust it with? What do they absolutely not trust it with? How do they think it might show up in their field?
The responses were all over the place, in the best way. This is a class made up of students coming from different corners of media, communication, design, and creative work. Some are thinking about sports media. Some are thinking about marketing, advertising, or PR. Some are coming from photography, film, video, UX, law, entrepreneurship, or other fields where AI is already starting to change what work looks like.
AI is not landing on all of them in the same way. For one student, it is a study tool. For another, it is a threat to creative client work. For someone else, it is already part of a business idea. For someone else, it is something they use but still feel weird about using.
What stood out to me was that most of the fear was not really about the technology itself. It was about what the technology might take away.
A lot of them were circling around authorship. If AI helps make something, is it still mine? If I use it to brainstorm, outline, revise, generate an image, or clean up a sentence, where does my work begin and end? At what point does help become replacement?
There was also a lot of concern about dependence. If AI does it for me, what am I learning? If it can write the first draft, make the image, fix the code, polish the sentence, or give me the answer, what happens to the part of me that was supposed to learn how to do that?
Another big thread was realness. Students talked about AI looking fake, sounding fake, giving wrong answers, making up information, or producing work that has that weird too-smooth feeling. Sometimes it is obvious: bad hands, strange lighting, generic writing, fake sources. But sometimes the harder question is what happens when nothing is technically wrong and it still feels off.
And then there was the jobs question. Especially entry-level jobs. If AI can do the basic version of a task, what happens to the people who were supposed to learn by doing those basic tasks?
That one feels especially prominent for students who are close to entering creative and media fields. They are not asking that question in the abstract. They are asking it because they can already see AI showing up in ads, headshots, social posts, video, design, writing, and client work.
So I took those reflections and brought the patterns back to class.
I didn’t want to walk in and say, “Here is what you all really mean.” That would defeat the whole point. I wanted them to hear their own concerns back in a slightly organized way, and then push deeper.
I put the concerns on the board in plain language:
Losing authorship.
AI as replacement.
Fake, unreliable, misinformation.
Dependence.
What am I learning?
Entry-level jobs disappearing.
Then I asked them to take their reflections back out and look for where these ideas showed up in their own writing. Not where they used the exact same words, but where the concern was present. What were they actually worried about?
From there, the activity became less about “Is AI good or bad?” and more about asking better questions.
For authorship, I wanted to get them away from the easy version of the debate. So we started with photography. Who in the room takes photos? Do you shoot digital or film? If you shoot digital, are you less of a photographer because you are not developing film by hand? If you color correct, crop, dodge, burn, or use Photoshop, is the photo less yours? Are you less creative because Photoshop is doing the work for you?
That led us toward the bigger question: where does creativity begin?
Is creativity in the hand labor? The original idea? The choices during the process? The edits? The final judgment? The fact that you know what you wanted in the first place?
That felt like the first real opening.
Just because the tool changes the process, does that automatically mean the tool owns the work? No. And at the same time, it does not mean the tool is irrelevant either. The question is where the person is in the process. What did they decide? What did they notice? What did they bring?
For dependence, we talked about tools we already accept.
Spellcheck. GPS. Calculators. Templates. None of these automatically make us lazy. But they can change where our focus is. GPS is useful, but if you never pay attention to where you are, you may not actually learn the place. Spellcheck is useful, but it does not mean you should stop caring about language. A calculator is useful, but depending on the situation, maybe you still need to understand what the math is doing.
So the question became: when does a tool remove friction, and when does it remove learning?
That feels more useful than asking, “Is AI cheating?” because it makes us name what the learning actually is. Not all struggle is valuable. Some struggle is just inefficient. But some struggle is where judgment gets built. Meaningful friction is the big thing, but where does the value of friction in each specific situation get decided?
For the realness part, I started with writing.
I asked them who was using AI to write in some way. I made it clear this was not a trap. I was not trying to catch anyone. But the room immediately did that thing students do when they all know the answer but nobody wants to be the first person to say it out loud. Sheepish smiles. Side eyes. A little quiet laughing.
So I said, okay, how about you just tell me what the hallmarks of AI writing are?
They had answers right away.
The em dash. The “it’s not X, it’s Y” structure. Certain words that feel like AI has a frequent flyer card for them. Delve. Transformative. Landscape. And then we talked about the way AI writing often feels like it is trying to land the plane too neatly. It has this little “one more thing” move at the end, like it wants to make sure you know it completed the assignment. It is polished, but in a way that starts to feel too patterned.
We also talked briefly about the idea of linguistic watermarking, where certain patterns in word choice might help identify AI-generated writing. They had not heard much about that before, and I think it helped them see that “AI voice” is not just a vibe; there are actual patterns there.
That gave us a way into visual AI.
I showed them this cat.
Generated from the prompt “generate an image of a cat,” with no other specifications. I wanted the most vanilla, generic thing the tool would make when I removed myself from the decision-making.
They all immediately said it was generated.
When I asked how they knew, one student said, “Because everything you show us is generated,” which was honestly fair. I should probably mix that up a bit.
But when I pushed them to name what specifically made it look generated, it got harder. They started searching for tiny things that might be wrong, but a lot of what they noticed was not actually wrong. The cat looked good. Too good, maybe.
Eventually someone said it was just too nice-looking, like the uncanny valley.
That was the turn. It wasn’t fake because it had six paws or melted whiskers. It felt fake because it was too perfect. The same way AI writing can feel fake because every sentence is clean, every transition is smooth, every paragraph knows exactly what it is doing.
There is no weirdness. No unevenness. No mistakes. No human variance.
One student mentioned taste here, so I asked her what taste is, and she had a hard time putting it into words. Which makes sense, because taste is one of those things we use constantly but struggle to define. So we kept circling it.
Taste is knowing what fits. Knowing what matters. Knowing when something is too much. Knowing when something technically works but still feels wrong. That is a hard thing to teach directly, but it is a big concern they had. How do you keep AI from flattening taste?
Last was the biggest fear, jobs.
The entry-level question is real. If AI can do a lot of the basic first-pass work, then what happens to the people who were supposed to learn by doing that work? So, I asked them: what is the difference between a junior-level person and a senior-level person?
At first, they did what students do and gave the obvious answer.
“The word junior.”
Ha. Ha. Ha.
But we kept going. Someone said a senior-level person leads a team. I asked why. Why do they get to lead the team?
Eventually we got to something more useful: they know what to do. They know what matters. They know what to include and what to leave out. They know when something is good enough, when it is wrong, when it is risky, when it needs more work, and when to stop.
That was the connection I wanted them to see.
A senior-level person is not valuable only because they can perform a task. They are valuable because they have judgment built from experience. They have context. They have taste. They can tell good from bad in situations where there may not be one obvious answer.
And that pulled the whole discussion together.
Their fears about authorship, dependence, fake work, misinformation, and jobs were not separate fears. They were all pointing at the same thing: losing the parts of the process where human judgment gets built.
That’s the through line.
The thing underneath all of these fears was not really the tool. It was the fear of losing the parts of the process that come from being a person. Lived experience. Knowledge. Taste. Context. Memory. Preference. Judgment. The weird collection of things you know because you are you.
That last part feels simple, but it is the whole point.
If you want to define yourself apart from AI, the word is already sitting there: you.
You are the difference. Not in some inspirational poster way. In a practical way.
AI does not know what you meant. It does not know what matters to you. It does not know the room, the audience, the client, the memory, the thing you noticed, the reason something feels off, or the reason something is worth doing in the first place.
So the question for this class is not going to be “Do we use AI or not?” That is too easy, and honestly, too late.
The better question is: where are you in the process?
On one end, you let AI do all of it. On the other end, you do all of it. Most real use cases are somewhere in the middle. Every prompt, revision, edit, approval, rejection, and final decision moves you somewhere on that line.
Not just whether AI was used, but where they were in the work. Where did their judgment show up? Where did their taste show up? Where did their experience show up? Where did they make a decision the tool could not make for them?
That’s the line we are going to keep coming back to. That’s what I want students to start seeing.
Because if the fear is replacement, the work is learning how to stay present in the process.