I want to be upfront about something before I write another word about AI: I use it. In my work, in my writing process, in building this site. I find it genuinely useful, and I'm not interested in performing skepticism I don't actually feel. But I also come to this conversation as someone with a background in political science and education, which means I have a hard time separating the technology from the conditions in which it is being deployed, who is making the decisions about it, and who is absorbing the costs.
The AI conversation, particularly in the creative and content industries, has a tendency to collapse into one of two stories. In the first, AI is a democratizing force, lowering the barriers to creation, enabling the solo founder and the independent artist to punch above their weight. In the second, it is an extinction event for human creativity, a race to the bottom that will hollow out the livelihoods of writers, illustrators, musicians, and everyone else who built a career on the value of human imagination. Both of these stories contain real observations. Neither is complete. And the choice between them tends to map uncomfortably onto who is telling it and what they stand to gain.
The labor question that keeps getting skipped
When industry observers talk about AI and creative work, they tend to focus on quality and productivity. Can AI write a good blog post? How much faster can a designer work with generative tools? These are real questions, but they are not the most important ones. The more important question is structural: what happens to the people whose livelihoods depended on the tasks that AI has now made cheap or free?
This is not a hypothetical. A 2023 survey by the Graphic Artists Guild found that a majority of professional illustrators reported losing work directly attributable to clients switching to AI image generation. Copywriters and content writers report similar patterns. Translators, voice actors, stock photographers: the list of creative workers whose market has been materially disrupted in the past two years is long and growing. The fact that new kinds of work may emerge to replace them is genuinely true and genuinely cold comfort to someone who spent a decade building expertise that is now being undercut by a model trained, in many cases, on their own work without compensation.
The training data question is one the industry has largely tried to talk around. The major AI companies built their models on internet-scale datasets that included, without meaningful consent or compensation, the creative work of millions of individuals. Writers, artists, and musicians are currently in various stages of litigation over this. The legal outcomes are uncertain; the ethical picture is less so. Benefiting from a tool while declining to examine how it was built is a choice, and it is worth naming it as one.
What "democratization" actually looks like in practice
The democratization argument for AI is worth taking seriously, because it contains real truth. The cost of producing professional-quality visual content, writing, and design has dropped significantly. For small businesses, nonprofits, community organizations, and individual creators who could never afford agency rates, this matters. There are people building things today that they genuinely could not have built two years ago, and that is a real good in the world.
But democratization is always worth examining more carefully, because the word tends to obscure as much as it reveals. Access to generative AI tools is not evenly distributed. The most capable models are behind paywalls or require technical knowledge to use effectively. The businesses and individuals who can afford to integrate AI into sophisticated workflows are, by and large, the ones who already had resources. The freelance writer with two decades of craft is losing work to a mid-sized company's AI subscription, not to another independent creator who finally got the tools to compete.
There's also a question about what kind of creative culture we're building. When the marginal cost of producing content approaches zero, the volume of content increases dramatically. We are already seeing this: the internet in 2026 is measurably noisier than it was in 2022. More content, less attention, and a growing premium on work that is visibly and credibly human. Whether that premium accrues broadly to human creators or narrowly to the most prominent and well-resourced ones is an open question, but the history of previous technological disruptions suggests the latter is more likely without deliberate intervention.
The education angle
My background in education gives me a particular angle on one part of this conversation that I think gets less attention than it deserves. AI is being integrated into educational contexts at a pace that has substantially outrun our capacity to understand its effects on learning. Students at every level are using AI to complete assignments, draft essays, generate code, and solve problems. Educators are scrambling to figure out what this means for assessment, for academic integrity, and for the actual development of skills.
The optimistic version of this story is that AI handles the rote and mechanical, freeing students to focus on higher-order thinking. The pessimistic version is that the struggle of working through a difficult problem, the frustration of a bad first draft, the iterative process of developing genuine competence, is precisely what builds the cognitive architecture that makes someone a capable thinker. If AI smooths over all of that friction, we may be producing people who can direct AI tools effectively while having quietly hollowed out the capacity for independent reasoning that would let them know when the AI is wrong.
I don't have a clean resolution to this. I don't think anyone does. But it seems important to hold the question openly rather than closing it off with either a celebration of efficiency or a blanket prohibition on the tools.
How I actually think about using it
Since this is a personal essay as much as a professional one, I'll say directly how I've landed on using AI in my own work. I use it as a research and drafting assistant, not as an author. The thinking, the framing, the voice, the editorial judgment about what to include and what to leave out: those are mine, and protecting them feels important not just for quality reasons but for reasons of integrity. I don't want to produce work that presents a perspective I haven't actually formed.
I also think about the broader context of what I'm contributing to. When I produce content using AI tools, I try to ask whether I'm adding something to the information environment that justifies its existence: a real perspective, a useful synthesis, a specific insight that serves the reader. The answer isn't always yes. Sometimes the right response is to produce less rather than more.
And I'm genuinely uncertain about some of the bigger questions. I think the training data issue is a real ethical problem that hasn't been adequately addressed. I think the labor displacement effects are serious and deserve policy responses that the industry has been successful at delaying. I think the concentration of AI capability in a small number of very large companies is a structural problem for competition, for governance, and for the diversity of the tools available to people who aren't their primary customers.
"Being thoughtful about AI doesn't mean refusing to use it. It means refusing to use it thoughtlessly."
What a more honest conversation sounds like
A more honest conversation about AI and creative work would hold several things at once. It would acknowledge genuine utility without treating it as sufficient justification for ignoring harm. It would name the workers who are being displaced without pretending that the technology can simply be disinvented. It would ask who is making the decisions about how this technology develops, and whether those decision-makers have adequate accountability to the people most affected by those decisions. And it would resist the pressure, which is real and commercial, to have a simple story.
For those of us who work in content, strategy, and digital communication, the honest position is probably something like this: we are operating in a moment of genuine disruption, some of which is generative and some of which is destructive, and the distribution of those effects is not neutral. Using these tools thoughtfully means staying in that complexity rather than resolving it prematurely in either direction.
That's not a tidy conclusion. It's not meant to be.