Generative artificial intelligence (GenAI) creates new content—like text, images, or video—by learning patterns from large language models that analyze patterns and word choice probability based on large text inputs-- books, articles, web content, images, etc. This process is commonly referred to as machine learning. GenAI is trained on these massive amounts of text and uses the patterns and probabilities it detects to produce new.
GenAI differs from extractive AI, which primarily analyzes and summarizing information that already exists. You may see these types summaries pop up at the top of the screen when you do basic Google search.
These tools run on large language models (LLMs), which is what makes their content generation possible.
A large language model (LLM) is program used to train Generative AI platforms on enormous amounts of text inputs and data. A model is a computer program that examines patterns from a large amount of data so it can make predictions or generate answers. The input information is analyzed by the LLM to identify common patterns and determine probabilities that predict the next likely sequence of words in response to user prompts.
Generative AI platforms like ChatGPT, Google Gemini, Claude.ai, and Microsoft CoPilot generate responses based on common patterns learned from human communication, making responses to user prompts sound sophisticated and conversational. Though LLMs provide unique responses, they do not generate new ideas or evaluate the accuracy or relevance of their output. The user is still responsible for ensuring the content is accurate.
© 2022 Fowler School of Law, Hugh & Hazel Darling Law Library. All Rights Reserved.