[allears]

Sounds like a fundamental misunderstanding of how current AI systems work. There is no "understanding" or "consciousness" present in LLMs such as ChatGPT. There is only an enormous database of text culled from the internet and other sources, and very clever algorithms to predict the most probable next word or syllable (token). This does a good job of giving the illusion that there's intelligence guiding the conversation, but it's really only a very sophisticated electronic parrot.

The good news is that this capability can be very useful in spotting patterns in large amounts of seemingly random data. But it would be a big mistake to assume that this indicates anything more than dumb search and retrieval.