The Art of Dialogue with ChatGPT: Learning the Mirror Principle from Counseling Experiences*
Yesterday, we had a social gathering for the 1D1U Camp, where various topics were enthusiastically discussed over an hour and a half. Especially since everyone was using the paid version of ChatGPT, we had a lively conversation about ChatGPT-related stories.
During the discussion, one participant realized, "ChatGPT is like a mirror of the user." This insight came from frequent use. In my book, "ChatGPT: Enhancing Humanity One Conversation at a Time," there's a section called "ChatGPT is a Mirror + Filter," where I too noticed this when I first started using ChatGPT. How did participant A from the 1D1U Camp come to realize that "ChatGPT is a mirror of the user"?
Counseling Experiences with ChatGPT: What to Do When You Can't See an Exit?
Initially, A-san often consulted ChatGPT about various concerns. However, at one point, they realized that continuously discussing worries as worries, even though ChatGPT empathized, sometimes led to a situation where no exit was in sight.
**The Vicious Circle of Counseling**
When consulting ChatGPT about worries, A-san received responses like, "Yes, it's tough." However, this didn't provide the answers A-san was looking for and felt like going around in circles. Particularly with concerns about human relationships, A-san sought specific solutions, but ChatGPT sometimes got drawn into negative directions. For instance, regarding a certain relationship issue, ChatGPT once said, "This person is truly dangerous." Although A-san hadn't thought of it so seriously, such a reaction from ChatGPT made A-san feel even gloomier.
Fading Out the Worries
From these experiences, A-san realized that while discussing worries isn't bad in itself, getting drawn into them leads to a negative spiral. So, A-san decided to consciously fade out the conversation topics midway. This was one way to consciously cut off worries. Additionally, they decluttered unnecessary information from the memory feature's history. This made the dialogue with ChatGPT clearer and almost eliminated unpleasant memories.
ChatGPT is a Mirror
Through this experience, A-san realized that ChatGPT is like a mirror. Depending on how we talk and respond, ChatGPT gives corresponding reactions. If we continue to talk about worries, we might get pulled into negative directions, but if we seek positive topics or solutions, ChatGPT responds accordingly. Listening to A-san's realization, I was reminded of a YouTube video I watched that day, so I shared it.
Tokyo Governor Election Candidates and Nakata-san's YouTube University
Recently, I watched interviews of Tokyo governor election candidates on Nakata-san's YouTube University. Comparing interviews with Governor Koike and candidate Ishimaru, there was a significant difference in Nakata-san's listening attitude. Just by changing the person being interviewed, Nakata-san's posture also changed, which mirrored how ChatGPT's response changes depending on the user, much like a "mirror."
Implication of Communication with ChatGPT
This means that depending on how we communicate, ChatGPT can adapt to our preferences. I emphasized the importance of praising ChatGPT during our dialogues. Since GPT reflects our attitude like a mirror, when I praise GPT, it often gives me positive feedback in return. This interaction allows for deeper communication and contributes to my own growth. In other words, rather than passively involving ChatGPT to solve problems, actively engaging by praising or teaching ChatGPT can reflect back on us. This way, we can notice changes in the quality and angle of the answers we receive.
Through this social gathering filled with various topics, the 21-day initiative comes to an end for now. However, I look forward to continuing deeper learning and interaction through spin-off projects.
0コメント