Technology
When NOT to Use AI (A List Nobody Wants to Write)
Everyone talks about when to use AI. Here is the harder list: when AI makes the work weaker, riskier, or less yours.

I use AI every day.
So no, this is not a purity sermon.
See an AI workflow guide and which AI tools are actually useful for freelancers.
But the harder skill is knowing when not to use AI.
Because the tool is so easy now that it can become a reflex.
And reflex is where bad work begins.
When You Need To Learn The Thing
If you need to explain it later, do not outsource the thinking now.
Use AI too early and you get an answer you cannot defend.
It may even be a good answer.
Still useless if you do not understand why.
The rule: if I need to explain it, defend it, teach it, or build on it, I do the first thinking myself.
AI can challenge it after.
AI can clean it after.
AI does not get the first swing.
When The Relationship Matters More Than The Output
Some messages need your fingerprints.
Hard feedback.
Apologies.
Sensitive client emails.
Condolence notes.
Anything where the other person would care that you actually wrote it.
This is not about text quality. AI can make a neat sentence.
It is about trust.
If a client finds out you used AI to write the apology for missing a deadline, the problem is not the wording.
The problem is that you made the repair feel outsourced.
I keep a hard boundary: anything emotionally loaded gets written by me. Period.
When You Are Dodging The Hard Part
This is the sneaky one.
AI can make avoidance feel productive.
Need to make a hard call?
Ask for five frameworks.
Need to talk to a client?
Draft the message twelve times.
Need to start the article?
Generate twenty angles and hate all of them.
Very busy. Very little courage.
The test: if you keep prompting because you do not want to decide, stop prompting.
Go do the hard human part.
I wrote about this same stall pattern in how to build momentum when everything feels stuck. The block is rarely just the words.
When Being Wrong Has A Real Cost
AI can be wrong with excellent posture.
That is the danger.
Wrong plus confident is worse than messy plus unsure.
For legal, medical, tax, contract, compliance, safety, or money decisions, do not let AI sit in the driver’s seat.
Use verified sources.
Use a qualified person when the stakes demand it.
Use AI only as a drafting or question-making tool if you still verify every important claim.
If checking the answer takes as long as doing the work properly, you did not save time.
You just moved the risk around.
When The Weird Take Is The Value
AI is very good at the average shape of things.
That is useful for many jobs.
It is dangerous when the weird edge is the whole value.
The strange angle.
The uncomfortable sentence.
The point you are not fully sure how to say yet.
Those are fragile.
If you ask AI too early, it will often sand them into something clean and dead.
The move: write the ugly human version first.
Then use AI for pressure-testing, structure, or missing angles.
Do not let it flatten the part only you would have noticed.
When The Client Hired Your Judgment
Some clients hire you for speed.
Some hire you for execution.
Some hire you for taste, voice, strategy, or judgment.
Know which one it is.
There is a difference between “AI helped me work faster” and “AI did the work and I put my name on it.”
That line may shift by client and project.
But do not pretend it does not exist.
I don’t have a universal rule here. But I have a test: if the client asked “did you use AI for this?” would I feel comfortable saying yes? If not, I probably shouldn’t have.
The Point Is Control
The goal is not to avoid AI.
The goal is to stay in charge.
Use it for drafts, questions, summaries, cleanup, checklists, and speed.
Keep the judgment, the relationship, the learning, and the weird take.
That is the line.
Cross it often enough and your work may still look fine.
It just will not feel like yours.
Advertisement