"AI" as in LLM (Large Language Models) is to be avoided whenever you need something to be accurate, or correct.
LLMs are statistical beasts, which can only ever output stuff that "looks very plausible" because it's constructed from stuff the LLM has seen before. It impresses humans because we like the plausible output generated with little effort on our behalf - we ignore the massive effort needed to generate the LLM model in the first place.
This "AI" has no intelligence, and no understanding of anything. It merely takes prompts in, and generates plausible-looking output based on the statistical weightings in its massive vector database. Any apparent intelligence or understanding is completely fake!
"AI" as in LLM (Large Language Models) is to be avoided whenever you need something to be accurate, or correct.
LLMs are statistical beasts, which can only ever output stuff that "looks very plausible" because it's constructed from stuff the LLM has seen before. It impresses humans because we like the plausible output generated with little effort on our behalf - we ignore the massive effort needed to generate the LLM model in the first place.
This "AI" has no intelligence, and no understanding of anything. It merely takes prompts in, and generates plausible-looking output based on the statistical weightings in its massive vector database. Any apparent intelligence or understanding is completely fake!