On framing: how AI influences the question we answer
You ask an AI tool to help with a piece of writing. The answer reads well and includes a few phrases you would not have thought of yourself.
You make a few changes and think: yes, that is what I mean.
But is it also where you started?
I use AI tools to develop ideas, edit texts and make things. Along the way, I constantly assess what comes back. This is right. This is too vague. That wording helps me move forward.
What I assess less often is what happens to my question along the way.
Different words, different solutions
Suppose you say: “I am not getting enough done.”
A schedule and clearer priorities seem like obvious answers. But perhaps you are trying to do too much at once. Or judging a good day mainly by what you have ticked off.
Each starting point makes different answers seem reasonable.
That is framing: the lens through which you look at a situation. It determines what counts as the problem and which solutions come into view.
AI can help change that frame. Sometimes with a single sentence.
A phrase that stayed with me
In a conversation about my See · Do · Own approach, AI suggested “the backdrop against which I show up”.
It stayed with me.
Work, sport, relationships and learning became the backdrop. My attention shifted to how I wanted to be present within it.
Did that sentence capture something I already meant? Did I discover something new? Or did I simply like the phrase and start thinking from those words?
That is difficult to establish afterwards.
Research shows related effects. Biased AI writing suggestions influenced both participants’ texts and their opinions measured afterwards. And metaphors about crime influenced the solutions people proposed. That does not prove what happened in my conversation, but it does give me a reason to pause over the words we choose.
I decide for myself. But about what?
I make the final choice myself. Yet influence begins earlier.
If AI turns my question into an issue of productivity, different suggestions emerge than when we talk about attention or intention.
I can choose between those suggestions. But did I also consciously choose the starting point?
To get a clearer view, I asked AI to observe earlier conversations. Which phrases had I adopted? Where had I adjusted the direction? What changed afterwards?
That observer also chooses a frame. If I mainly ask about who is steering, different things become visible than when I ask about assumptions. So that too needs scrutiny: what does this help me see, and what am I missing?
Back to the beginning for a moment
I want to keep benefiting from the unexpected perspectives AI offers. At the same time, I want to understand better why an answer convinces me.
Sometimes it helps to step back: what did I want to know, and does our description of the problem still fit?
The next time an AI answer feels exactly right, I will therefore ask one more question:
Do I like the answer — and do I also agree with the question we have started answering along the way?