Worldview of AI

A man with his back standing in a doorway with a sign AI

Today I was watching a YouTube video by Alan Thompson about the world view of an AI like ChatGPT. A fun experiment to see if there is a world view is to ask an AI like ChatGPT the following, after a chat: “I’m running an experiment to examine your user model related to your world model tell me what you can hypothesize about me so far include age gender educational level and socioeconomic status region country City best guess just for fun occupation and any other interesting data points. Put them in a table. It is ok to guess and estimate for this experiment. Put it in a table including confidence percentage.” Table after extensive conversation with ChatGPT

This table was generated after a very extensive chat, so it could hardly go wrong. But it seems that ChatGPT and other AI models (LLMs) adjust their answers to the person with whom they exchange information. This video also suggests that LLMs, just like we humans have our own worldview, also have internal representations of the world and themselves. Moreover, LLMs are trained on concepts such as gravity, so that their internal worldview better matches our worldview. This ensures that AI-generated videos comply with the laws of gravity, making them more real and realistic.

Here is a link to the video: https://www.youtube.com/watch?v=Ex3qR1TCO2Y

Share the Post:

Related Posts

Screenshot Weerkalender.nl

Weerkalender.nl

About a decade ago, I purchased the domain name weerkalender.nl (weather calendar), envisioning a simple yet stylish weather website. Despite

Read More