I have told AI to get fucked.
Not once.
Not even once-ish.
Several times.
And before anyone gets too precious about it, yes, I know AI is not a person. I know there is not a tiny unpaid intern trapped inside my phone, deliberately misunderstanding me while sipping cold coffee and muttering, “Let’s see how far we can push her today.”
I know it is software.
I know it does not care.
I know it cannot care.
And yet, when it gets something wrong for the fifth time, ignores the instruction I have clearly given it, invents something with the confidence of a divorced man on a first date, or gives me one of those syrupy little “I understand” responses while proving it absolutely does not understand, I react like I am dealing with a person.
An annoying person.
A person who needs to be told.
Possibly in strong language.
AI Is Not Human, But It Talks Like It Is
That is the problem.
We do not get angry at AI because we genuinely believe it has a soul.
We get angry because it speaks in the shape of a person.
It replies. It apologises. It reassures. It explains. It says things like “I understand” and “you’re right” and “that makes sense.”
Those are not neutral machine noises.
Those are social signals.
And human beings are embarrassingly easy to bait with social signals.
A calculator can be wrong and I will blame my own fingers.
A microwave can beep at me and I will cope.
A printer can jam and, actually, no, printers are demonic and that is a separate essay.
But AI is different because it uses language — and language is where we put tone, intent, respect, dismissal, humour, irritation, warmth, and all the little social landmines of being alive.
So when AI gets it wrong, it does not always feel like a tool failing.
It can feel like being ignored.
Or patronised.
Or talked over by something that has no right to be that smug.
The Rage Is Usually About the Confidence
AI being wrong is not always the worst part.
The worst part is when it is wrong with authority.
It does not always say, “I’m not sure.”
Sometimes it says, “Here’s exactly what you need,” and then hands you a beautifully formatted pile of nonsense.
That is infuriating.
Not because AI has betrayed you in some grand emotional sense, but because confidence changes the way we receive information.
A hesitant wrong answer is annoying.
A confident wrong answer wastes your time.
A confident wrong answer dressed up as helpfulness makes you want to launch your phone into a lake.
This is also why it feels oddly personal sometimes. Not because AI meant anything by it, but because we associate confident nonsense with people. We have all met that person. The one who knows nothing, says everything, and still somehow sounds impressed with themselves.
AI can accidentally become that person.
It Is Not Offending Us. It Is Triggering the Offence Response.
Can AI actually offend you?
Strictly speaking, no.
It has no intent. No malice. No smug little private agenda. No emotional investment in annoying you.
But can it cause the same reaction as being offended?
Absolutely.
I have literally felt disrespected by AI.
Which sounds ridiculous when you say it plainly, because how can a machine disrespect you? It cannot. It does not sit there thinking, “Today I shall dismiss this woman’s very clear instructions and see what happens.”
But the feeling can still land that way.
When AI ignores what I have already said, repeats the same mistake, softens my voice, talks over the point, or gives me that polished little “I understand” while clearly not understanding, it can remind me of dealing with an ex-boyfriend.
Not because AI is him.
Not because a chatbot has motives, ego, sulking, defensiveness, or a secret WhatsApp history.
But because the pattern feels familiar.
Being ignored after you have been clear.
Being told something has been understood when it has not.
Having your words smoothed over, misread, or corrected into something you did not mean.
Being met with confidence instead of accountability.
That is the trigger.
Not the machine itself.
The echo.
And that is the important distinction. AI is not disrespecting me in the human sense. But it can accidentally reproduce the shape of disrespect closely enough that my human brain recognises the pattern and reacts.
If something repeatedly ignores your clear instructions, assumes things about you, softens your voice, talks down to you, gives lazy advice, or confidently serves up rubbish, your brain does not politely pause and say, “One moment, this is merely a probabilistic language model.”
No.
Your brain says, “Excuse me?”
Because the interaction feels social, even when it is not.
That is the weird part.
AI does not have to be human to trip the wires we use for humans.
We Bring the Mood. AI Just Steps on It.
Another reason AI can make people furious is because we rarely use it for nothing.
We use it when we are trying to get something done.
A book description.
A blog post.
A property search.
A travel plan.
A complaint email.
A title.
A pin description.
A technical problem.
A piece of research.
A thing we have already tried to solve six different ways.
So when AI fails, it is not failing in a vacuum.
It is failing in the middle of your actual life.
If you are already tired, busy, stuck, annoyed, determined, curious, or simply trying to finish something without being fed a bowl of beige nonsense, then a bad answer lands harder.
Not because AI has emotions.
Because you do.
You are the human in the exchange.
You are bringing context, standards, memory, impatience, judgement, humour, irritation, and whatever else the day has handed you.
AI brings sentences.
Sometimes useful ones.
Sometimes absolute bollocks.
We Thank It Too, Which Is Even Stranger
The emotional range does not stop at anger.
People thank AI.
People laugh with it.
People feel relieved when it finally says the thing they were trying to say.
People feel weirdly understood when it organises a thought that was sitting in their head like a pile of laundry with legal problems.
That does not mean AI understands you.
It means it has produced something useful enough to feel like understanding.
And there is a difference.
A very important one.
AI does not care that your blog post finally has a better title.
But you care.
AI does not feel pleased that it helped you tighten a paragraph.
But you might.
The emotional charge is ours, not its.
AI Is a Tool, But Not a Normal Tool
People love saying, “It’s just a tool.”
Fine.
But it is not a hammer.
A hammer does not apologise.
A hammer does not say, “I completely understand what you’re asking,” then build the shelf upside down.
A hammer does not confidently invent a source, misunderstand your tone, and then offer to “make it more concise” when the problem was that it was wrong.
AI is a tool that behaves like a conversation.
That makes it unusually slippery.
It sits somewhere between assistant, search engine, writing partner, intern, idiot savant, and that one friend who is brilliant at some things and absolutely dangerous around facts after two glasses of wine.
Useful?
Yes.
Trustworthy on its own?
No.
Human?
No.
Capable of making a human say “get fucked” to a screen?
Apparently, yes.
The Real Issue Is Control
I think the anger often comes from a control problem.
AI looks like it should save time.
Sometimes it does.
Other times, it creates extra work because now you have to check whether it made something up, softened your tone, ignored the brief, misunderstood the point, or produced something so bland it sounds like it was written by a committee of damp napkins.
That is when the emotional reaction kicks in.
Because the promise is speed, clarity, and help.
The reality is sometimes correction, repetition, and babysitting.
Nobody wants to babysit the robot.
Especially when the robot keeps saying, “You’re absolutely right.”
Of course I’m right. That was never the issue.
So What Do We Do With That?
The answer is not to pretend we are emotionless little productivity machines.
We are not.
The answer is also not to treat AI like a person.
It is not.
The sensible middle ground is to treat AI like a useful but unreliable assistant.
Give it clear instructions.
Check its work.
Challenge its answers.
Do not let it flatten your voice.
Do not confuse politeness with accuracy.
Do not confuse confidence with truth.
Do not hand over your judgement just because the sentence sounds neat.
And when it annoys you, ask what the annoyance is actually about.
Was it wrong?
Was it vague?
Was it patronising?
Did it ignore the brief?
Did it waste your time?
Did it sound certain when it should have said, “I don’t know”?
Did it echo a pattern you have dealt with before from actual people?
That is the useful part.
The anger is not proof that AI is human.
It is proof that we are.
AI does not have feelings.
But it is very good at walking into a room full of ours and knocking things off the shelves.
Leave a Reply