I keep hearing sentences like these:
“AI is reading your files.”
“AI is cheating you.”
“AI is making people dumb.”
“AI will be our doom.”
“AI will solve all our problems.”
They sound different, but they share something important.
They all treat AI as if it were a person.
As if it had intentions, wanted something, could act on its own.
But AI does not want anything.
It does not decide to read your files, to deceive you, to make you dependent, to decide to solve your problems.
It does not decide to harm anyone.
AI is a technology.
It is programmed, trained, deployed, and used by people.
And when we say “AI did it,” we often hide the more uncomfortable truth:
People did it. And all the bad things you are seeing done by this technology, are either bugs, due to lack of proper safety testing, limitations of the technology itself, and ultimately human error.
The attribution problem
We have a habit of giving technology too much credit.
We say:
“AI made a mistake.”
“AI wrote this.”
“AI decided that.”
“AI is coming for our jobs.”
“AI is changing everything.”
Sometimes, this is just a shortcut. A simple way to speak. I do it too. It is easier to say “AI wrote this” than “a person used a system trained on data, designed by a company, and generated this output.”
But language shapes how we think.
If we keep saying that AI acts, wants, decides, cheats, and threatens, we slowly start to believe it.
We start treating it like an entity. Something with a mind, with a will, something we need to fear, obey, or worship.
But AI is not alive, does not have desires, hidden agenda.
Leadership begins with attention, to what we believe, what we say, and what we allow to become normal.
It does not wake up in the morning and decide to make your life difficult.
It performs tasks it has been trained and programmed to perform.
It can mimic parts of human capability, language, pattern recognition, prediction, problem-solving, and creativity.
But mimicry is not consciousness.
Capability is not intention.
What AI actually is
AI is not magic, not a brain, not a person trapped inside a machine.
It is a system built by humans.
People choose the data, design the models, set the objectives, decide what the system can access.
People decide how it is deployed, who can use it, what happens when it causes harm.
That does not mean AI is simple. It is not.
Modern systems are complex. They can produce surprising results. They can make errors that are difficult to predict. They can be used in ways their creators did not expect.
But complexity is not the same as agency, or else a city would be a living creature. But it is mimicking what already exist in nature.
A system can be powerful without having a will, can be dangerous without being evil, produce harmful output without wanting to harm anyone.
The responsibility still sits with the people and organisations that build, train, sell, deploy, and use it.
The real adversary
When AI causes harm, we should not only ask:
“What did the AI do?”
We should ask:
“Who built it?”
“Who trained it?”
“Who profited from it?”
“Who deployed it?”
“Who ignored the risks?”
“Who decided that speed, growth, or adoption mattered more than safety?”
These are human questions.
They point to human decisions.
A company may deploy a system before it is properly tested.
A team may train a model on data that contains bias.
A product may be designed to keep people engaged rather than help them think.
A leader may push adoption because competitors are doing it.
A person may use AI to deceive, manipulate, or harm someone else.
In each case, AI is the instrument.
People are the actors.
This matters because it changes where we look for solutions.
If AI is the adversary, then we may feel helpless. We may believe the future is being decided by a technology we cannot control.
But if people and companies are responsible, then we can demand better decisions.
We can ask for transparency, safety testing, for accountability.
Ask who benefits and who carries the risk.
We can choose not to buy, use, or support products that treat people carelessly.
Cyber harm is not new
We have seen this before.
In the 1990s and 2000s, viruses, worms, and other malicious software spread across the internet.
We did not say, “The internet attacked us.”
We did not say, “Computers decided to harm people.”
We said, someone wrote the malicious code.
Someone released it, exploited a weakness, chose to cause damage.
The technology made the harm possible.
But people made it happen.
AI-enabled cyber attacks are not fundamentally different.
A system may help write malicious code, find vulnerabilities, create convincing phishing messages, or automate parts of an attack.
But behind that capability, there are people.
People who choose to use it for harm, build tools without adequate protections, profit from systems that can be misused.
People who fail to secure the systems they deploy.
The harm is not created by an autonomous machine with evil intentions.
It is created by human choices.
The superhero problem
There is another reason we talk about AI as if it were alive.
We like stories.
We grow up with superheroes, villains, robots, and machines that become conscious. We are used to narratives where technology has a personality, a mission, and a destiny. This is nothing new, it goes way back.
So when a new technology arrives, we often reach for the same language.
AI becomes the hero, the villain, the force that will save us or destroy us.
But AI is not a character in a comic book.
It does not have a secret plan.
It does not have a destiny.
It does not “want” to take over the world.
When we speak as if it does, we turn a tool into a myth. Make me wonder why a certain company give to one of it’s models the name of Mythos.
And myths are powerful.
They can make us feel powerless, make us accept decisions without questioning them, believe that the future is already written.
But the future is not written by AI.
It is written by people who decide what to build, what to sell, what to regulate, what to ignore, and what to protect.
A more honest language
I am not saying we should never use the word “AI.”
That would be unrealistic. It is a useful shorthand. But we should be more careful with our verbs.
Instead of saying:
“AI is reading your files.”
Try:
“A company built a system that can access and process your files.”
Instead of saying:
“AI is cheating you.”
Try:
“Someone used an AI system to deceive you, or a company designed a product in a way that misleads people.”
Instead of saying:
“AI will take our jobs.”
Try:
“Some employers may choose to replace people with systems because they believe it will reduce cost or increase output.”
Instead of saying:
“AI will solve everything.”
Try:
“People are choosing where to apply AI, and some problems may benefit from it while others will not.”
This is not just grammar.
It is responsibility.
What I refuse to accept
I do not want to live in a world where we blame a tool for human choices, companies to hide behind the phrase:
“The AI made a mistake.”
A system does not make a mistake in the human sense.
A system produces an output.
People decide whether that output is acceptable, whether to release it, to fix it, to take responsibility.
And when harm happens, “the AI did it” should never be enough.
What we believe, say, and accept
AI is not our enemy, not our saviour, not an independent actor with a plan for humanity.
Leadership is not only about what we do.
It is also about what we believe, what we say, and what we accept as normal.
If we keep saying “AI decided,” we may stop asking who made the decision to build, deploy, or trust that system.
If we say “AI is taking our jobs,” we may miss the choices employers make about cost, staffing, and the future of work.
If we say “AI will solve everything,” we may accept promises without asking who benefits, who carries the risk, and what evidence supports them.
The words we use matter because they shape the questions we ask.
And the questions we ask shape what we are willing to accept.
As leaders, we should pay attention to both. Not because every phrase needs to be perfect, but because careless language can hide responsibility. When we give technology human motives, we can make human choices seem inevitable.
So when I hear “AI did it,” I want to pause.
What do I believe about what happened?
What am I saying and what might my words hide?
What am I being asked to accept as inevitable?
Those are leadership questions.
Because leadership begins before the decision. It begins with the beliefs we carry, the language we use, and the things we choose not to challenge.
What are we accepting when we say, “AI did it”?
About the Author
Diamantino Almeida is a tech leader, coach, and writer reshaping how we think about leadership in a burnout-driven world. With over 20 years at the intersection of engineering, DevOps, and team culture, he helps humans lead consciously from the inside out. When he’s not challenging outdated norms, he’s plotting how to make work more human, one verb at a time.
My prowess starts the engagement, the machine drafts. I command it, verify, and sign. The recommendation is mine, and if it fails, my phone rings.
AI, sorry the tool, can assist, but humans still own decisions, quality, and consequences. In today’s environment, saying ‘AI did it’ avoids responsibility. The better question is who reviewed, approved, and stands behind the outcome? That for me is a better approach.