I once described an AI technology using the "divination" paradigm.

【新着連載】Over the AI ―― AIの向こう側に（9）：へつらう人工知能 ～巧みな質問を繰り返して心の中をのぞき見るhttps://t.co/1r6Jo2PznJ pic.twitter.com/HlebKvu0b7

? EE Times Japan編集部 (@eetimes_jp) March 30, 2017

In this column, I concluded,

(1) The business of "fortune-telling" is a collaborative creative activity based on a conversation between a "fortune-teller" and a "customer who is going for a fortune-telling session.

(2) In (1) above, the expected output and process "is to derive 'the answer that the customer doing the divination wants.

I am still absolutely confident in this conclusion, so if anyone has a problem with it, please dispute it with me.

When I see a self-proclaimed "fortune teller" speaking on a TV program, I immediately walk away from the TV.

It's because I feel like my eyes are dirty, my ears are rotten, and I've lost time (even if it's only for a few seconds).

However, I do not believe I have the right to "harm even the feelings of those who find these programs 'enjoyable' -- especially if they are my family members.

So, the right thing to do in these situations is to "walk away quietly.

------

On page 5 of the column introduced in this issue, I have compiled a table of past incidents that led to the "near extinction of the human race.

And, well, all of this is human error and has nothing to do with AI technology.

Recently, however, I have felt that "AI technology" has caught up with me to the point where I cannot ignore it.

------

As you know, I have been using ChatGPT and GitHub Copilot a lot lately.

And the way it is used is just like a Copilot.

By offering advice, receiving a response, and then engaging in further discussion of some of the content, getting a response, and so forth, and so on and on...

"Thoughts lurking inside I emerge as verbalized/embodied."

I can achieve the above.

Two important things are,

(1) AI technology is not the main focus; it is my partner.

And,

(2) The ability to keep "repeating" as many times as possible.

ChatGPT and GitHub Copilot "never get tired" (I'm the only one who gets tired), no matter how long I use them.

Of course, ChatGPT and GitHub Copilot can also be self-contradictory (or "circular argument") in their arguments after 20 or more times of poking and prodding.

This self-contradiction and circular argument seems (at the moment) to be something that ChatGPT and GitHub Copilot cannot even notice on their own.

It is up to me, the human being, to notice this and change the stage of the discussion by changing the content of the question or the background of the setting -- for now.

-----

What I am trying to say is,

"AI" cannot destroy humanity, but it is increasingly likely that "humans" will find approaches to destroy humankind in discussions with AI."

This is what I 'feel'.

What is essential is that the above phrase is not something I derived by "logic" but something I actually "feel" as I am confronted with AI daily.

-----

Frankly, this reality is quite "uncomfortable" for me.

Why do I find it uncomfortable? -- It is because the relationship between the "generated AI (ChatGPT and GitHub Copilot)" and "I" is the same as the relationship between a "fortune teller" and a "fortune teller's customer," which I make fun of.

Of course, I believe that fortune-telling is incomparably more "vulgar" in its character and significance.

However, even if we limit ourselves to the relationship and approach, I think it is not too much to say that the two are not "similar" but rather "joint."

The only difference is that the object of belief is either a "fortune teller" or a "generative AI."

This fact now makes me feel intensely uncomfortable.