The contestable opposite is the industry default: most developers use AI invisibly because clients didn’t ask, and the work ships unmarked. I won’t do that.
What I do
- Document AI use in every handoff. If a code module was generated, audited, or refined with an AI assistant, the documentation says so and names the model.
- Show the human-review step. Generated output I ship passes through me first. The handoff makes that explicit.
- Price for my time, not the model’s. You pay $275/hr for senior judgement, not for tokens.
- Treat AI-assisted writing the same way. Drafts, audits, reviews — if AI helped, the byline reflects it.
What I decline
- Hidden AI use, including in code I deliver, copy I write, or images I generate.
- Charging for AI-generated output as if it were hand-crafted from a blank page.
- “Ghostwriting with AI” engagements where the deliverable is dressed up to look like something else.
- Pretending I don’t use AI. I do, every day, openly.
Why this is the position
I’m proud of how I use AI. It’s a tool that lets me deliver more for less time and audit my own work harder than a solo developer could otherwise. Hiding the tool would suggest I’m ashamed of using it. I’m not. You should know which parts of your project benefited from AI assistance, which parts were my hand alone, and that I reviewed every line either way.
Disclosure is also where the buying market is going. Edu and gov are already asking. I’d rather have a real answer than invent one under pressure.
See also
- Delivery engagements are fixed price. Same idea, different surface — buyers deserve to know what they're buying before they sign.
- I teach you how I built it. That's part of the build.. On the disclosure that happens after the work ships — handoff that doesn't hide what was used.
- What you owe the people still running your old code. On the longer obligation that disclosure is part of — to the next developer who picks up the file.
