In this article
AI is a useful tool: research, variations, time saved. It is not an idea, not a point of view, not a method. And confusing the two is the surest way to wreck a creative career.
I will say it plainly, at the risk of displeasing: I am against dependency on AI in creative work. Not against AI. Against dependency.
My position, plainly
Yes, it is a useful tool for research, variations, saving time on thankless tasks. No, it is not an idea. Not a point of view. Not a method. And an excellent prompter without methodology is still a designer without methodology.
The four reasons
- Real skills versus prompts. When I hire, a small technical test is enough: fundamental gaps in composition, hierarchy, software command and UX show up immediately. AI does not mask the absence of method, it exposes it.
- Efficiency. I do not pay a collaborator to wait for a result. I pay for an approach, argued choices, a reusable graphic system.
- Client trust and legal risk. Many clients refuse AI in deliverables. The murkiness around intellectual property and sources remains a real risk. We deliver work that is clear, traceable and controlled.
- Production files. AI can produce an image. It still cannot deliver a usable production file, with clean paths, variants and print constraints.
AI can produce. Creating means thinking, framing, iterating, owning. A creative's value is vision and responsibility for the choices.
Creating is not producing
That is the whole distinction. Production is execution, it automates. Creation is a chain of decisions someone must answer for.
There is also a dependency issue, exactly like SaaS products depending on their servers: when your skill rests entirely on a tool, you do not have a skill, you have a subscription. The day the tool changes its rules, its pricing or its output, you are left with nothing.
This is not a principled stand against progress. It is the observation that the market will end up paying for what the machine cannot do: decide and own it.
Frequently asked questions
-
Is this a rejection of AI?
- No, a rejection of dependency. AI is useful for research, variations and saving time. It replaces neither method nor decision.
-
Why do some clients refuse AI in deliverables?
- Because of murkiness around intellectual property and sources. A traceable, controlled deliverable is a commercial argument, not a constraint.
-
Doesn't a designer who masters AI have an edge?
- Yes, provided the method sits underneath. AI amplifies what exists: a solid approach, or its absence.