I'm currently in my undergraduate degree and I have been studying AI ethics under one of my professors for a while. I always have been a partisan of Searle's strong AI and I never really found the chinese room argument compelling.
Personally I found that the systems argument against the chinese room to make a lot of sense. My first time reading "Minds, Brains, and Programs" I thought Searle's rebuttal was not very well structured and I found it a little logically incorrect. He mentions that if you take away the room and allow the person to internalize all the things inside the system, that he still will not have understanding--and that no part of the system can have understanding since he is the entire system.
I always was confused on why he cannot have understanding, since I imagine this kind of language theatrics is very similar to how we communicate; I couldn't understand how this means artificial intelligence cannot have true understanding.
Now on another read I was able to draw some parallels to Nigel Richards--the man who won the french scrabble championship by memorizing the french dictionary. I havent seen anyone talk about this online so I just want to propose a few questions:
- Does Nigel Richards have an understanding of the french language ?
- Does Nigel serve as a de facto chinese room ?
- What is different between Nigel's understanding of the french language compared to a native speaker?
- Do you think that this is similar to how people accredit LLMs' to simple prediction machines?
- And finally, would an LLM have a better or worse understanding of language in comparison to Nigel?
- What does this mean when it comes to the our ideas of consciousness? Do we humanize the idea of thinking too much when maybe (like the example) we are more similar to LLMs than previously thought?
[link] [comments]