Over time, I believe that, thanks to AI, the majority of my work will be focused on creative endeavor.
In that light, I asked Claude AI to help me become more creatively autonomous.
Here’s part of what it suggests:
01 Creative brief generation
Use AI to turn a vague impulse into a specific, scoped brief—with constraints, form, audience, and a single clear intent.
02 Taste calibration
Feed AI examples of work you love and hate. Have it articulate your aesthetic in precise language. This gives you a vocabulary for your own taste.
03 Rapid prototyping of ideas
Generate 10-20 divergent versions of an idea—different angles, tones, formats, or audiences. Not to use them, but to find which direction pulls you.
04 Structured creative review
Use AI as a first reader—not for praise, but for honest diagnosis. Ask it to identify where you work loses momentum, where intent and execution diverge, or where the reader might get confused.
05 Influence mapping
Describe your work to AI and have it surface artists, writers, movements, or techniques you likely haven’t encountered. Then have it explain what specifically connects them to your sensibility.
06 Constraint design
Ask AI to design creative constraints tailored to your specific creative block or tendency. If you over-explain, it gives you a word limit. If you avoid emotion, it bans abstraction.
07 Audience pressure-testing
Have AI roleplay as specific audience members—a skeptic, an expert, a casual reader, a hostile critic. Put your work in front of them before it’s public.
08 Creative process documentation
Use AI to help you narrate your process in real time—decisions made, paths abandoned, breakthroughs. Over time this becomes a personal creative methodology you can teach, refine, and share. Your process becomes your intellectual property.
#01 — Creative brief generation is the right place to start. Every other action on this list presupposes you know what you’re making. It’s also the most transferable skill across creative disciplines.
It’s also reversible. A brief is low-stakes to generate and discard, which means you can demonstrate the full loop — impulse → AI collaboration → usable brief → actual work — without needing months of output to show for it.
Questions worth sitting with:
- What’s the medium or domain where you feel most creatively blocked right now — writing, visual work, a project, something else?
- Is there a part of your creative autonomy you’re worried AI might undermine rather than support? That tension is often the most instructive place to start.
I then asked Claude to explore creative brief generation as my first (public) creative autonomy case study.
Here’s part of what happened.
First, it asked me what creative domain I’m working in or most want to develop (visual art / design). Then, it asked me where my creative process most often stalls (I finish but don’t know what’s working). Finally, it asked me what my relationship is to creative briefs right now (I genuinely don’t know what a good brief looks like for my work).
Here’s what it said:
Questions worth sitting with:
- When you finish a piece and feel uncertain — is the doubt more “I don’t know if this is technically good” or “I don’t know if this is mine“? Those require very different briefs.
- If a brief tells you what success looks like before you start, is there a risk it makes you stop trusting the instincts that emerge during making? Where’s the line between a brief that guides and one that over-determines?
Clause then created a brief generator tool that empowers me to
• describe my impulse, idea, or feeling—as raw as it is
• choose a visual-based medium / format (with an “Undecided—help me choose” option)
• primary audience
• answer “What does ‘not working’ look like for you?”
Clause then offered a few things worth noting and narrating publicly:
The “failure” field is the most important one. By asking you to describe a past piece that left you uncertain, the AI can reverse-engineer what your brief was missing — not what a generic brief should contain. Your uncertainty is diagnostic data.
The success condition it generates is deliberately in-process, not retrospective. Most artists evaluate work after it’s done; this brief trains you to know what “working” looks like while you’re still inside it. That’s the shift from uncertainty to autonomy.
The autonomy note in the output is the one to watch over time. After five or six briefs, patterns will emerge — what your briefs consistently avoid naming, what constraints you resist, what questions your work keeps circling. That pattern is your creative methodology starting to become visible.
A brief can over-determine. The protection against that is in the wording of the success condition — it should describe a quality of experience, not a visual outcome. “This works when I feel resistance to editing it” is a brief. “This works when the composition is balanced” is a specification. The difference matters.