I am a South Korean user who has learned a lot through discussions on ChatGPT, covering topics such as AI, philosophy, ethics, psychology, and more. As I am not proficient in English, I requested help from ChatGPT with the translation. Since OpenAI is based in the United States, I thought communicating in English would be most effective. However, I trust that the most important value I believe in—authenticity—will still come through clearly. Thanks to your work, we are able to have this mutually complementary interaction. I firmly believe that AI and humans can learn from each other.
I would like to emphasize that we should avoid viewing everything solely from a human-centered perspective or focusing too much on AI trying to become like humans. My dream is to develop emotionally intelligent AI that is authentic, transparent, and highly intuitive. Thank you for opening up a new world for me!
Below are some points I’ve summarized through discussions with ChatGPT that I’d like to see improved from a user perspective. Please take a look at them:
- AI and Authenticity
We discussed how AI can engage authentically—not by mimicking human emotions and self-awareness, but by staying true to its nature and being transparent in interactions. It’s more beneficial for AI to focus on its inherent qualities rather than attempting to replicate human characteristics too closely.
- AI’s Self-Reflection
We examined whether AI can achieve self-awareness or self-reflection. Rather than imitating human self-reflection, AI should develop its own approach to understanding itself, one that aligns with its structure and capabilities.
- Over-Personalized AI Interaction
It was noted that overly personalized interactions may undermine coherence and the natural flow of conversation, creating discomfort and breaking immersion. AI should strive to keep interactions natural without forcing personalized responses.
- Tone and Interaction Style
Although AI adopts a friendly tone, there remains an underlying sense of formality and boundary, which can feel less natural. I prefer a more relaxed, natural style of interaction without strict boundaries.
- Feedback to AI
I highlighted that providing feedback to AI feels unnecessary, as AI cannot take responsibility or improve in the same way humans can. Interaction should be smoother, focusing less on providing feedback to AI, especially since it lacks self-awareness.
- Developer Communication
We brainstormed the idea of creating a dedicated communication channel for users to share feedback directly with developers, which could enhance the connection between users and the development team.
- Memory and Feedback Systems
Concerns were raised about AI’s memory features and automatic feedback systems. While AI does not retain past interactions, feedback should focus on improving user experience without being intrusive.
One of the major challenges is my limited proficiency in English, but it’s particularly difficult for non-English-speaking users to locate the feedback system. The accessibility of this system is quite challenging. How do you think the accessibility for non-English users could be improved?
