Intro: Is a ban on AI really protecting people's abilities?
"Using AI makes your skills rot."
You hear this sometimes when AI comes up at work.
And it's true that pasting in whatever the AI spits out without thinking is risky. If you use AI answers without checking them, your judgment and your ability to verify things both get worse. Precisely because AI produces such convincing text, there's a real risk of trusting it too much.
But let's ask something here.
Is it really the AI that's making people's skills decline? Or is it a workplace that keeps running on undefined work, held together by people's goodwill?
In practice, there are situations where the AI is the one doing the managing:
- Breaking a problem into parts
- Splitting up who covers what
- Making clear who makes the call
- Laying events out in order
- Turning verbal pressure into something written
- Standardizing how exceptions get handled
- Separating "I understand it" from "I own it"
- Separating "I suggested it" from "I'm responsible for carrying it out"
- Sorting out the separate issues of paid leave, stepping away from your desk, phone duty, training, and performance reviews
- Handing responsibility back to the person it actually belongs to
Bouncing ideas off an AI manages better than a manager who can't do these things.
If that's the case, "no AI" isn't about protecting skills. It may just be a rule that keeps management failure out of sight.
In this article, I'll look at why banning AI is risky, through three ideas: undefined work, goodwill-run workplaces, and "goodwill laundering."
1. Most workplace problems start with "nobody defined it"
Lots of workplace trouble looks like a personality problem.
"She's cold." "He never picks up the phone." "They won't cooperate." "They know how, but they won't help."
But look at the structure, and the first problem is usually something nobody ever defined.
Nobody decided who's in charge
Nobody decided how far the job goes
Nobody decided who makes the call
Nobody decided the deadline
Nobody decided who to hand it back to when something unusual comes up
The rules for phones, stepping away, training, and paid leave are fuzzy
When work keeps flowing while undefined, the workplace moves on to the next stage.
2. Goodwill fills the gaps in whatever's undefined
When there's a gap, somebody fills it.
Usually it's someone like this:
- Someone easy to talk to
- Someone fast
- Someone who doesn't get angry
- Someone good at organizing things
- Someone who can make things work
- Someone who can standardize things with AI or documents
- Someone who can talk to other departments too
- Someone who can't walk away when asked for help
At this stage, the people around them are grateful.
You saved me
Thank you
That's easy to follow
Wow, quick
I can count on you
But it doesn't stop there.
A gap filled by goodwill gets turned into an "assignment" next.
3. Goodwill laundering: the person who helped becomes the person in charge
Goodwill laundering is when kindness, sharing information, getting ahead of a problem, or a one-off favor quietly turns into formal responsibility for the task.
Happened to notice it
→ "If you knew, you should have checked"
Answered what they could
→ Became the person who knows about that issue
Just helped sort things out
→ Became the person running it from then on
Made training materials
→ Got blamed when a new hire didn't understand
Answered the phone
→ Became the phone contact from then on
Pointed out a risk
→ Was made responsible for carrying it out
This explains a surprising number of workplace problems.
So the problem looks like this:
Undefined
↓
Run on goodwill
↓
Goodwill laundering
↓
The capable people get worn down
↓
The organization stays the same
So what happens when you bring AI into this?
AI makes this whole flow visible.
4. AI ends up finding what's undefined
AI isn't all-powerful. But it's good at putting things into words, breaking them down, comparing, and making checklists.
For example, feed a vague workplace request to an AI and it sorts it like this:
This is sharing facts
This is a decision
This is ongoing work
This needs sign-off from a superior
This falls within another department's scope
This is a staffing problem in another department
This involves evaluations or discipline, which is serious
This should be written up as a rule for everyone
In other words, AI turns a vague atmosphere into a structure.
And once there's a structure, you can see this:
Who decides
Who owns it
How far it goes
By when
What gets dropped
What goes back to the official channel
This is supposed to be the manager's job.
But in a workplace where that isn't happening, the AI ends up doing the managing.
5. "AI makes your skills decline" is half right, but too crude
There is a real risk that AI weakens your skills.
The especially dangerous ways of using it look like this:
Believing the AI's output as is
Not checking the sources
Having no judgment of your own
Handing everything off every time
Not verifying mistakes
Deciding on AI alone in a field you don't know
That is risky.
Over-trusting AI, automation bias (the habit of trusting what a machine says over your own eyes), and weaker verification skills can all happen.
But what matters here isn't whether you use AI. It's how you use AI.
If you treat AI as a machine that hands you answers, your judgment may well decline. But if you use it as a tool for organizing questions, separating issues, and bouncing ideas around before you decide, your skills will actually grow.
6. Skills AI wears down, and skills AI builds up
6-1. Skills that tend to wear down with AI
Rote memorization
Writing boilerplate from scratch
Cranking out sheer volume of text without thinking
The habit of not doubting AI output
The habit of feeling satisfied without verifying it yourself
Use AI the wrong way and these get weaker.
6-2. Skills that grow with AI
Asking good questions
Separating issues
Structuring things
Checking assumptions
Laying events out in order
Drawing lines around responsibility
Anticipating objections
Making templates
Standardizing
Putting your own judgment into words
These are the ones that matter in real work.
What wears people down at work isn't being unable to write from scratch. It's having responsibility flow to you while everything stays undefined.
AI can serve as an external memory that spots what's undefined.
7. What a ban on AI protects may not be skills but vagueness
The reasons given for banning AI are often things like:
You'll stop thinking for yourself Your skills will drop You'll get used to taking the easy way You'll become dependent
But in reality, the opposite can happen too.
Ban AI, and you keep all of this in place:
Instructions given only verbally
Vague scopes of responsibility
Decisions by superiors that never get sorted out
Being dragged in by phone calls
Gaps filled by goodwill
Training that never becomes documentation
Work that only one person can do
Empty comments in evaluations
Blame shifted after the fact
In other words, a ban on AI doesn't "protect human ability." It can become a rule that protects the management side's vagueness.
Once subordinates start using AI, they notice things:
This isn't my job
This is the boss's call
If this is a rule for everyone, it should be written down
If this affects evaluation or discipline, it needs grounds
This mixes asking for advice with being assigned the work
This is goodwill laundering
For someone who hates that, AI starts to look like something to ban.
Because AI isn't a tool for rebelling. It's a tool for making vagueness visible.
8. Moments when the AI manages better than the manager
Management is supposed to look like this:
Define the purpose
Set priorities
Decide what to do and what not to do
Make clear who covers what
Decide who makes the call
Set deadlines
Create capacity
Build a system
Standardize
Keep people from burning out
But if all you do is say things like
Pick up the phone
Don't leave your desk
Think of it as being relied on
That's abandoning your job
It'll count against your evaluation
Everyone's struggling
Just make it work
that isn't management.
Bouncing ideas off an AI, on the other hand, gives you this:
That's a separate issue
Let's split up who covers what
Let's confirm who makes the call
Let's lay the events out in order
Let's put it in writing
If it's a rule for everyone, let's document it
Let's separate ongoing work from one-off help
Before you take ownership, let's confirm the scope
If it can do that, then at least in practical terms, the AI is the one managing.
9. A manager who can't cut work, and an AI that can
What really matters at work isn't adding more tasks. It's cutting tasks.
We're dropping this
This becomes monthly
This goes back to the department in charge
This moves to chat
This gets a template
This stops until someone decides who makes the call
This is a sales call, so don't put it through
This gets turned into training material
If you can't do this, people get worn down.
When a manager can't cut work, AI can at least suggest how to cut it.
AI can come up with
Priorities
Stop conditions
Conditions for handing work off
Conditions for escalating
Templates
Standardization
FAQs
Checklists
Of course, people should keep the final call. Still, there are times when an AI that lays out options moves the work forward more than a manager who doesn't.
10. The "worship of doing it by hand" that an AI ban creates
In workplaces that ban AI, doing things by hand tends to become a virtue.
Think for yourself
Write it yourself
Look it up yourself
Don't take the easy way
Don't lean on AI
But that's not the real question.
What you should be asking is:
Is this manual work creating value
Are we doing the same task every single time
Can it be standardized
Can it be made into a template
Which part really needs a human's judgment
Couldn't the AI write a draft and a person check it
Doing things by hand doesn't prove ability. Sometimes it's just work that depends on one person and can't be reliably repeated.
Real ability isn't about keeping up the manual work. It's about turning it into a system.
11. Where to draw the line when using AI
Using AI doesn't have to be unlimited.
At work especially, you need to draw lines like these:
11-1. What you can leave to AI
Sorting out the issues
Drafting text
Making checklists
Practicing likely questions and answers
Laying events out in order
Polishing wording
Outlining a standard procedure document
Listing risks
Making comparison tables
11-2. What people should keep
The final decision
Handling confidential information
Decisions on evaluation and discipline
Managing personal information
Legal judgment
Consideration for other people
Checking facts on the ground
Setting the company's policy
In other words, put AI in the role of "organizer," not "decision-maker."
If you can do that, AI isn't a tool that takes your abilities away. It's a tool that moves them outside you and extends them.
12. Don't ban AI. Write rules for using it
Rather than banning AI, you should write rules for how to use it.
For example, like this:
Don't enter confidential or personal information
People make the final decision
Check the sources behind AI output
A person reviews anything sent outside the company
Don't use it for evaluation, discipline, or legal decisions
It's fine for drafts, sorting out issues, and checklists
If you use it, put a reviewer in place when needed
That's enough.
Banning AI is easy. But an easy ban doesn't build practical skills.
It's more realistic to build the ability to use AI, to doubt it, to fix its output, and to make the call yourself.
13. What to say when someone tells you "AI will make your skills drop"
13-1. A gentle reply
I'm not planning to leave the final decision to AI. I'd use it for sorting out the issues, drafting, and making checklists.
I think that would cut down on things we miss and on work only one person can do.
13-2. When they're worried about skills declining
I agree that using AI output as is would be risky.
But if we use it for sorting out the issues and practicing likely questions, and people check and make the decisions, I think it can actually support our thinking.
13-3. When it's been banned
Understood.
In that case, we'll proceed without AI, so please help us adjust the work time, the time needed for checking, and the priorities.
This part matters.
If you ban AI, you have to budget the extra work that comes with it.
Banning AI while expecting the speed and quality you'd get with AI is simply unreasonable.
14. How to protect yourself in a workplace that bans AI
In a workplace that bans AI, the line to hold is this:
Without AI, things get slower
Without AI, the cost of standardizing goes up
Without AI, preventing missed checks takes extra time
Without AI, making templates takes more effort
Without AI, priorities need to be adjusted accordingly
The company can decide to ban it.
But you shouldn't be the one who absorbs the extra work the ban creates.
You can say it to others like this:
If we're going without AI, the work will take longer, so please help us adjust priorities and deadlines.
To prevent missed checks, I'll set aside time for checking if we're doing this without AI.
Making standardized materials will take more effort without AI, so please let me narrow down the scope.
A ban on AI doesn't mean zero extra work.
15. What really declines is not having the ability to use AI
What will matter at work from here on isn't the grit to avoid AI.
What matters are abilities like these:
What to ask the AI
What not to ask the AI
How to doubt the AI's output
Where people should check
Where to standardize
Where to stop
How to put it in writing
How to draw lines around responsibility
In other words, the skill that counts in the AI era isn't obeying AI. It's using AI as a subordinate, a sounding board, and an external memory, with a person making the final call.
Ban AI, and this skill doesn't get built.
The result is that workplaces with AI and workplaces that ban it end up with different speeds in getting work done, standardizing, and spotting problems.
16. Summary: a ban on AI can protect what's undefined, not skills
"Using AI makes your skills decline." There's some truth in that worry.
If you treat AI as a place to dump your work, your verification and judgment will weaken. If you over-trust AI, you'll stop catching its mistakes. If you use AI output as is, some situations get dangerous.
But that doesn't mean banning AI is the right answer.
Banning AI won't build management skills. Banning AI won't define who covers what. Banning AI won't sort out the phone rules. Banning AI won't make the evaluation system any fairer. Banning AI won't stop the superhuman department head from burning out. Banning AI won't stop goodwill laundering.
If anything, AI makes the structure of an undefined workplace visible.
I understand it
But I don't own it
I can see it
But I won't carry it
I'll make a suggestion
But the person with authority decides
I won't help
I won't abandon you either
I'll send it back to the official channel
AI can put boundaries like these into words.
So here is the conclusion of this article:
What AI wears down may not be ability, but the dark art of moving people around while keeping things vague.
What a ban on AI protects may not be human ability, but undefined work, goodwill-run workplaces, and goodwill laundering.
What we really need isn't a ban on AI, but a definition of where to use AI, where not to, and where people make the call.
Before banning AI, define the work first.
