“Can you summarise this article for me?”
“What’s the key point of this meeting?”
“Give me the use cases from this document.”
“Explain this concept in simple terms.”
I catch myself asking AI these kinds of questions all the time now.
Sometimes it is genuinely helpful. Sometimes it is just the quickest way to get through a lot of information.
But I have started to wonder what happens when I skip too quickly to the AI version of something. If AI is doing more of the summarising for me, how much of the thinking am I still doing myself, and then is the perspective really mine?
When the shortcut becomes the default
This came up in a team meeting recently. We were talking about how quickly AI has moved from something interesting on the side to something many of us now rely on without much thought.
We are stretched across work, family, study and deadlines, so the shortcut makes sense:
“I’ll get AI to summarise the chapter; pull out the key concepts; explain the use cases.”
The part of the conversation during this meeting that stayed with me was the statement: “If I only ever read the AI summarised version, I might feel informed, but I may not have really understood anything for myself or built my own perspective”.
An AI summary is never completely neutral. Even when it is useful and accurate, it has still made choices. It has decided what to keep, what to leave out, and what to make sound important.
That is not always a bad thing. But the danger is that it can smooth out the parts where you would normally think: the uneasy sentence, the contradiction, the odd detail, the bit you disagree with but keep thinking about afterwards. Those are usually the places where your own perspective is grown from.
When efficiency comes at the cost of perspective
We lose the ability to form our own perspective
If a tool has already decided what matters, we begin with its interpretation before we have had the chance to develop our own.
That framing can shape how we understand the rest of the information, even when we are not conscious of it.
We outsource the thinking, but the thinking is where we get our perspective
For me, learning usually happens when I reread something, get frustrated because I do not understand it yet, or realise I do not agree with what is being said.
It is not always efficient, but that struggle is often where the idea starts to become mine.
When AI removes all of that friction, it may also remove part of the process through which we develop judgement, curiosity and critical thinking.
We confuse having the information with having a view of our own
AI can sort, structure and tidy information very well. But it does not bring my context to it.
It does not know why something matters in this team, in this project, or at this point in my study.
It cannot fully account for what I value, what I have experienced or what I am trying to achieve.
That judgement still has to come from me.
So where does that leave us?
My view is not that we should stop using AI. I do not think that would be realistic, nor useful.
However, there is a difference between using AI to get oriented and letting it decide what something means.
In a busy week, sometimes I do need help finding the shape of an idea before I can properly engage with it. But after that, I think I still need to go back to the parts that matter and read them myself.
If AI gives me a summary, I want to ask what has been left out. If it helps me form a first view, I want to test that view rather than simply adopt it.
There may also be a better way to use AI. Once I have done some thinking of my own, it can be useful to ask AI what I might be missing, what assumptions I am making or how someone else might see the issue.
That feels very different from asking it to think on my behalf.
It is more like asking it to push back on me and challenge my thinking.
Protecting our ability to think critically
For me, perspective is not just the conclusion I arrive at. It is the route I took to get there.
It is shaped by what I have seen, what I value, what I am unsure about and the questions I have learned to ask.
If AI does too much of the interpreting for me, I worry that I will slowly stop building that muscle. And that muscle matters.
In a team, a project, or even an ordinary conversation, the useful thing is rarely that someone can repeat the summary.
It is that they have sat with the problem long enough to have a view that sounds like it belongs to them.
