On my habit of avoiding reading "below the final separator line" when discussing with generative AI

(Approx. 2000 characters)

It has been a while since I started using generative AI, but
I recently noticed a certain habit of mine.

When I want to have a discussion with generative AI,
I have found myself not reading to the end of the response.
To be precise, I have consciously stopped reading the
"Why don't you try this next?"
or
"How about considering this perspective?" sections that suggest the next steps written below the final separator line.

I didn't decide to do this from the beginning.
It is more like a feeling I had after realizing,
"Huh, I haven't been reading this part lately."

Looking back, I have come to think that this is not just a reading habit, but a matter related to
the very stance of my thinking.

Why did I stop reading "below the separator line" of generative AI?

Generative AI's responses are very kind.
They organize the points, present premises, list possibilities,
and finally show a path by saying, "Then, let's do this next."

However, when I read that final suggestion,
a certain change would occur within me.

Even if I am not fully convinced by the conclusion,
the "theme to think about next"
and the "direction to dig deeper"
unconsciously lean toward the direction suggested by the AI.

Why does the AI's "next suggestion" pull my thinking?

This is not to say that the AI is bad.
On the contrary, the AI's suggestions are very well-crafted.

• Logical

• Organized

• Easy to agree with

That is precisely why people naturally go along with them.

The problem is the timing of encountering that suggestion.

As a result,

The act of not reading below the separator line,
I believe, was a way to stop this flow once.

Not reading 'below the final separator' is not an abandonment of thought

I am not rejecting information from AI.

Theabovepart of the final separator, in other words

• organizing the points

• clarifying the premises

• presenting perspectives

I actually read actively.

However,

I do not read those parts.

I stop once, and ask myself,
'Having read this far, what have I thought?'

Thatmarginwas necessary for me.

Why do I write subjects and predicates like in English?

Around the same time, I noticed another habit.
When discussing with generative AI,
I had unconsciously startedwriting subjects, predicates, objects, etc., clearly.

In Japanese, conversations can function even when the subject is omitted.
It is convenient, but because of that,

In English,
you must explicitly state the subject of your thoughts, such as
I think / I assume / I question.

When you establish your stance by saying 'I think this way' before engaging in dialogue,
you become able to receive the AI's organization and supplementary informationas assistance.

Is this something unique to AI?

Come to think of it, this is not limited to generative AI.
The same thing often happens in discussions between humans.

When talking to someone who is a good speaker and has a well-organized logic,
the discussion can sometimes proceed within the other person's framework of understanding and organization
without your own questions being sufficiently shared.
(Be especially careful with people who are good at debating.)

As a result,
'what the point of the discussion was supposed to be'
and 'what parts were still undetermined'
become ambiguous along the way.

Generative AI is an entity that does this

That is why, even if your intent is slightly off,
a 'plausible-sounding discussion' can still take place.

It is the same with humans,
and the more I want to have a proper discussion,
the more I tend to speak by clarifying subjects and predicates, just like in English.

This is not so much to protect myself, but
to accurately convey to the other party 'what I am making the point of discussion'.

'I don't understand this part'
'I am not yet convinced by this premise'
'I only want to discuss this specific point'

If you don't state these clearly,
the discussion itself will drift elsewhere.

What is happening in dialogue with generative AI
might just be a reproduction of human discussion
in an extremely quiet and extremely orderly form.

Conclusion

When I discuss things with generative AI,
what I value is neither efficiency nor the correct answer.

It is all to ensure that,

The next time you interact with Generative AI,
if you have a little time to spare,
please try pausing once before the separator.

In that silence,
your own question might still be there.

Strawberry Dog🐶

#GenerativeAI #ControlOfThought #DialogueWithAI #AskingQuestions #WhatItMeansToThink #Verbalization #Introspection #TechnologyAndHumans

I also recommend this article💁

I would appreciate it if you could like and follow me☺️