The More You Use AI, the More You Need a 'Record of Judgment'

Previously, I wrote an article titled “Asking Questions Has Become My Job”.

Following that, I considered whether “decision-making” and “responsibility” are the jobs that will remain for humans in the AI era.

So, where does the record of human decision-making go?

Personally, after having AI generate multiple options, I often make the final decision on the spot.

The interaction with AI remains in the chat. However, my own reasoning for “why I chose this option” surprisingly does not remain.

Perhaps what becomes necessary as you use AI more is not the saving of answers, but the recording of judgments.

AI's answers remain, but your own reasons fade away

When you ask AI a question, multiple options appear in a short amount of time.

You can have it compare the pros and cons of each, and by providing additional conditions, you can narrow down the candidates further.

And in the end, you decide for yourself, “Let’s go with this one this time.”

The problem is that even if you are convinced at that moment, you haven't recorded the reason for your choice.

What did you prioritize?
Why didn't you adopt the other options?
Which information did you trust?
What did you leave uncertain?

These judgments cannot be understood just by looking at the finished text or materials.

Only the results remain, and the reasons that led to them disappear.

Chat history is not a record of judgment

You might think that saving your conversation with AI is enough.

However, what remains in the chat is what the AI proposed.

It does not necessarily record which option the human adopted and why they adopted it.

If you changed direction in the middle of a long exchange, it is even harder to understand.

AI response history and human judgment records are different things.

I think it is sometimes more useful to leave a few lines explaining the reasons for your final decision than to save every single proposal generated by AI.

There was a decision-making process behind choosing this article as well.

This time, I was torn about what the next article should be.

One option was to delve deeper into the responsibility of Governor Saito and politicians, which I covered in the previous article.

The other was to return to the original theme of AI and individual work.

There are reasons to pursue either path.

If I continue with the Hyogo Prefecture issue, I can think concretely about the responsibilities of politicians. On the other hand, if I stack political articles one after another, the original discussion on AI and work might become harder to see.

Therefore, this time, I decided to return to the AI side first and write an article considering the 'record of decisions'.

A simple record of this decision would look like the following.

Question
Should I cover the responsibility of politicians or the discussion on AI and work next?

Options
Delve into the Governor Saito issue / Return to AI and individual work.

Decision Made
Return to the AI work discussion first.

Reason for Decision
Because the starting point of this series of articles was the change in work due to AI, and I want to solidify that flow first.

What was deferred
The issue of Governor Saito and the third-party committee will be handled as a separate political series.

Even just this much allows me to understand 'why I chose this order' when looking back later.

A decision log only needs five items

A record of decisions does not mean you need to create long meeting minutes.

If you were to create detailed records for every task, it would become a burden in itself.

I think the following five items are sufficient to start with.

1. What are you trying to decide?

2. What options were available?

3. What was ultimately chosen?

4. Why was that chosen?

5. What was left undecided?

What is important is not summarizing the AI's response.

It is about recording what you prioritized and which uncertainties you accepted.

The purpose of recording is not to justify yourself.

A decision log is not meant to prove that you were right.

Rather, it is meant to make it easier to correct course when you make a mistake.

When the result is different from what you expected, if the decision-making materials from that time remain, you can think about where the problem lay.

Was there not enough information?
Were the priorities wrong?
Did you trust the AI's suggestion too much?
Did you underestimate the risks you were aware of?

If the decision-making process is not left behind, you will only look at the result and end up saying, 'It was a failure' or 'I'll be more careful next time.'

That makes it difficult to apply experience to the next decision.

Work in the AI era now includes the task of leaving behind traces of your decisions.

Thanks to AI, the time spent on research, organization, and writing has been reduced.

Because of that, humans will be making more decisions in less time than before.

That is precisely why it becomes necessary to record not just what you decided, but why you decided it.

Pose a question.
Get options from the AI.
Make a human decision.
Take responsibility for the result.
And leave the reason for the decision for next time.

Work in the AI era might just become this kind of flow.

I myself have not properly recorded the reasons for my decisions until now. I want to try starting by leaving a few lines behind, at least when I make major decisions, like how I chose the topic for this article.

Whether those records will truly be useful is something we will verify from here on out.

▼ Read also

・Asking questions has become my job

https://note.com/poliplus/n/n2f0ca1c8fe99

・In the AI era, what remains for humans is 'decision-making' and 'responsibility'

https://note.com/poliplus/n/n2e4c902fa319

・AI operational design for deploying a single video across multiple media—the decisions and responsibilities left to humans

https://note.com/poliplus/n/nfa84bc37af7a

・I tried keeping a decision log for just three days. The hardest part to write was the 'reason'

https://note.com/poliplus/n/n4d8b3698ef73

・What politicians need is not infallibility, but the willingness not to run away from mistakes

https://note.com/poliplus/n/nda124b166847

▼ Summary of this theme

・Summary of AI x communication know-how

https://note.com/poliplus/n/ncd6194c20727