New Year's Post (2026)

Full enclosure... I am not a fan of AI. OK - that is a bit of a generalization. There are some fantastic uses of AI Agents, and Generative AI can be fun. However, as a creative, I worry about the "AI Feedback Loop" and the narrowing of expression. As an educator, I worry about students using AI to get to the designation while skipping the journey altogether.

Instead of getting on my soapbox and decrying the evils of AI, I have decided to spend the year exploring AI. In fact, I started by asking ChatGPT for some ideas. The response was a bit long winded, but here are some challenges that I like (with some modification on my part).

Use AI to touch-up my photos
Use AI to generate summaries of my Obsidian notes
Use AI to scaffold a branding project
Build an AI-powered tool or workflow (or AI Agent)
Use AI to help design a problem-solving curriculum

My jump onto the AI bandwagon did not come out of no where. A educator friend of mine shared her excitement from a a course she took last year training AI on specific resources. A colleague recently built an AI Agent. Both were the last people I would have thought to be AI experts (or, at least more-than-novices).

So with their recommendations, and a bit of worry that I have become a luddite, I opened up my browser and created an account with Claude (Anthropic).

First Real Gen AI Experience

I have used AI to compose emails, produce mock data for an app, and get ideas on color schemes. But this wasn't the level of AI experience I was looking for. I wanted a collaborative experience where we worked together to produce some "deliverable".

The Prompt

I started by telling Claude that I wanted to design a problem-solving curriculum, and asked what details I should consider and what information I should include in the prompt to better assure its response addressed my goals. Little did I know what I was getting myself into.

I spent two hours chatting with Claude to better refine my goals and define how the curriculum would be structured. Most of the time was me reflecting on Claude's questions and doing further research, but it was a real back and forth. I was impressed with the way I was shepherded from a vary vague question to a well thought-out description of a curriculum.

A Journey

The process itself was an education. I got out of it exactly what I put in.

I could have easily had Claude make all of the decisions, define the goals, impose a structure, and generate a finished curriculum. That approach works, and it is tempting. Instead, I chose to engage with the questions it was asking. I followed up on its prompts, pushed back on a few recommendations, and found myself exploring ideas and learning models I hadn’t encountered before.

Rather than treating AI as something that simply delivers an answer, I began to treat it more like a guide, one that kept nudging me to clarify what I actually wanted. The value wasn’t just in the final curriculum, but in the thinking that led to it.

When I was teaching, I used to open my classroom early for students who wanted a place to hang out before first period. Some used the time to socialize. Others tried to finish homework. One student (I'll call him Rich) fell into the first group.

One morning, Rich came to me asking for help with his math homework. I was genuinely excited. He showed me the problems he hadn’t finished, and I started asking questions to understand what he knew and where he was stuck.

He stopped me and said, “I just need the answers.”

I so was disappointed. He wasn't interested in understand the concepts. He just wanted someone to finish his homework.

Years later, I see that moment differently. Rich had thirty minutes before class. He wasn’t choosing between learning and not learning. He was choosing between turning something in or not. He didn’t have the luxury of the long view.

AI is the latest resource that makes it easy to skip the journey and jump straight to the result. And sometimes, that’s exactly what people need. Constraints are real. Deadlines matter.

My concern is that Generative AI is being promoted almost entirely for its ability to produce deliverables: essays, code, images, math homework. And it does that remarkably well. What’s talked about far less is its potential to act as a guide, one that helps users clarify goals, explore alternatives, and wrestle with ideas they don’t yet fully understand.

Derivative Work?

One criticism I’ve heard repeatedly from friends and colleagues is that AI just repeats what has already been published. It doesn’t produce original, truly fresh, work. To test that claim, I asked ChatGPT to design an original UI feedback system for a web application. Instead of presenting a novel design, it responded by listing eight existing feedback patterns and suggested combining them into a hybrid system depending on the type of feedback.

Can a hybrid solution be called “original”? In this case, I would argue no. Each approach is immediately recognizable, and while the combination may be a useful solution, the originality lies more in the packaging than in the idea itself.

Copyright law gives us a helpful lens here: "A derivative work is a new creation that is based on one or more existing works… It must include significant changes to the original work to be eligible for copyright protection." This definition highlights an important distinction. Originality is not about assembling known components; it’s about transforming them in meaningful ways.

This is where many critiques of generative AI stop short. On the surface, AI tends to stay safely within the boundaries of existing patterns. It is very good at summarizing, remixing, and recombining what already exists. What it struggles with is deciding which assumptions to challenge, which constraints to remove, and which tradeoffs are worth making. Those decisions are where originality actually comes from.

Used as a guide rather than a generator, AI can help creatives move beyond surface-level hybrids. By asking it to explain why a pattern exists, what problem it was originally meant to solve, and under what conditions it fails, the tool becomes a way to interrogate existing ideas rather than simply reuse them. That process creates space for transformation.

Original solutions often emerge not from entirely new ingredients, but from changing the relationships between familiar ones. When a creator uses AI to explore alternatives, test assumptions, or deliberately push ideas into new territory, the result can move beyond derivative work. The AI doesn’t supply the originality. It supports the thinking that leads to it.

Seen this way, generative AI is less a creativity engine and more a catalyst. It can identify the raw ideas quickly, but it is the human who must decide what to discard, what to distort, and what to reimagine. The more intentional the journey, the more likely the destination will be something genuinely new.

On the Bandwagon?

Does this mean that I have drunk the Kool-Aid, and am now looking to AI to transform my life? No! I am still a skeptic. But my skepticism has changed shape.

AI in itself is not evil, nor is it inherently caustic to learning or creativity. However, it does make it easy to engage in shortcuts, bypass uncertainty, and arrive at something that looks finished. While this can be useful, it also makes it tempting to miss the parts of the process where learning and originality actually take place.

For students, AI can offer a way to get through an assignment without ever engaging with the ideas behind it. For creatives, it can narrow expression by steering work toward what already exists. These outcomes aren’t inevitable, but they are the certainly tempting.

This is the same criticism levied on the Internet, and there are many other new technologies that skeptics paned when introduced. As with these other technologies, we need to be aware of the potential for misuse, their impact on society, economy, and the environment. That said, I am excited to embark on my year exploring how AI can be used in my life.

For students, AI can become a way to get through an assignment without ever engaging the ideas behind it. For creatives, it can quietly narrow expression by steering work toward what already exists. Those outcomes aren’t inevitable—but they are the default if we treat AI primarily as a shortcut.

That’s why I’m choosing to spend this year exploring AI deliberately, not enthusiastically. I’m interested less in what it can produce for me, and more in how it can support thinking, questioning, and growth—without replacing them. Used carelessly, AI collapses the journey. Used intentionally, it can help illuminate it.