I Know I Shouldn't, But I Can't Help It

On AI, the blame game, and the person holding the tool

There is something deliciously convenient about blaming the tool.

AI made me do it.

The algorithm made me do it.

The chatbot said so.

It's a strange kind of innocence we're inventing for ourselves: grown people with keyboards, credit cards, careers, grudges, ambitions, and common sense — suddenly helpless the moment a machine answers back.

I don't buy it.

Artificial intelligence is not an oracle or a moral compass. It doesn't descend from the clouds carrying truth on stone tablets. It is an extraordinary tool still being shaped. And like every powerful tool humans have made, it comes with instructions which include: use with caution.

Responsibility falls not only on the creator but on the user too.

A knife can prepare dinner or save a life. It can also cause harm. The comparison is imperfect because AI is vastly more complicated, built by institutions, trained on human material, shaped by commercial incentives, safety rules, omissions, and the enormous mess of human history.

So no, I'm not arguing technology is neutral.

But neither are we.

That's the part that gets conveniently lost in the blame game.

We seem determined to have only two conversations about AI. In one, it's a miracle that will solve everything. In the other, it's a monster creeping toward the bedroom while humanity sleeps. Both stories let us avoid an uncomfortable third possibility: what if the machine is powerful, imperfect, still developing, and we remain responsible for what we ask it to do?

I mean often times I use AI. There, I said it. I use it to research, to organize, to challenge an idea, to locate what I've forgotten, to carry some of the administrative weight life keeps dropping at my door. That doesn't mean it does the work for me. It can't know what I meant unless I tell it, and sometimes even then it gets it wrong. It can't decide what I believe, can't take responsibility for what I publish, can't stand behind my name when the consequences arrive.

That part is mine.

If I ask AI to help me lie, manipulate, harass, plagiarize, or dress nonsense up as certainty, I don't become innocent because a machine helped with the grammar. If I publish something false because I couldn't be bothered to check it, ‘the AI said so’ isn't accountability, it's an alibi.

But the reverse matters too. We shouldn't treat every use of AI as evidence of laziness or intellectual collapse. Assistance is not surrender. People have always used tools to extend what they can do: calculators, spellcheck, search engines, mobility devices, maps, libraries, other people.

The moral question isn't simply whether a tool was used. The better questions are: What was it used for? Who checked the result? Who benefits, who could be harmed, and who is willing to be accountable when the answer is wrong?

That last question is where the teeth are — because we love responsibility right up until it points back at us. Companies blame users. Users blame companies. Platforms blame bad actors. Institutions hide behind policy. Everyone gets to point somewhere else while the damage sits in the middle, waiting for an owner.

No. Somebody still chose. Somebody still designed. Somebody still deployed. Somebody still clicked send.

AI deserves scrutiny. So do the people and institutions building it, and the systems deciding what it can see, what it rewards, and whose knowledge it treats as authoritative. And users deserve scrutiny too. That's not anti-AI, that's what taking AI seriously looks like.

I don't need a machine to be innocent before I use it. I need to understand that it's fallible, powerful, unfinished, and capable of amplifying both human brilliance and human foolishness. And I need to remember that I'm still in the room.

Maybe that's what makes this moment so uncomfortable: we want extraordinary tools without extraordinary responsibility.

I know I shouldn't, but I can't help it, I keep coming back to the human being holding the handle. Not because the tool carries no responsibility of its own.

Not because designers and corporations get a free pass.

Not because bias, exploitation, surveillance, misinformation, environmental cost, labour, or power disappear simply because a user typed the prompt.

But because accountability that stops everywhere except at ourselves isn't accountability at all.

Bringing me back to the point of rather than dismissing it, learn what it is, what it can do, where it can fail, and how to use it properly. Then decide what you think of it.

And in reminder of what my own grandma would have said, "In other words: don’t be a tool about the tool."

So use the tool. Question it. Argue with it. Verify it. Reject it. Make it better if you can.

Just don't hand it your conscience, and then blame it for what comes back.