Large language model (LLM)
Japanese 大規模言語モデル
A large language model is an AI trained on huge amounts of text to predict the next words, so it can write and answer in plain language. Chat-style AI tools such as ChatGPT run on one.
A large language model is a program that has read a very large amount of text and learned how words tend to follow one another. It builds a reply by choosing the most likely next piece of text, again and again. You can picture it as the autocomplete on your phone, scaled up enormously. Many different services can be built on top of the same model.
Watch out Because it writes what sounds likely, it can state wrong facts with confidence. Check numbers and dates against the original source, and read a service's terms before you paste in customer data or private documents.
Source: Google for Developers, Machine Learning Crash Course