Meet Generative AI

Flipped 0 of 5

• LLMLarge Language ModelFlip to revealLLMAn AI model trained on very large amounts of text so it can predict and generate language, one small piece (a token) at a time.Example: Chatbots such as ChatGPT, Gemini and Copilot are built on large language models.
• PromptYour instruction to the AIFlip to revealPromptThe question, text or instruction you give an AI tool. Its wording, context and detail shape what the AI produces.Example: "Explain photosynthesis to a first-year student in three sentences" gives a more focused answer than "photosynthesis".
• Training DataWhat the AI learned fromFlip to revealTraining DataThe large collection of text, images or other content used to train an AI model. The model learns patterns from this data, including its gaps and errors.Example: If the training data includes few sources about a topic or culture, outputs about it may be thin or inaccurate.
• HallucinationConfident, but wrongFlip to revealHallucinationWhen an AI produces information that sounds convincing but is false or made up, such as a fake quote, statistic or reference.Example: An AI lists a journal article with real-looking authors and page numbers, but the article does not exist.
• BiasUnfair patterns in outputsFlip to revealBiasWhen AI outputs reflect unfair assumptions or stereotypes, often learned from patterns in the training data.Example: If an image generator asked for "a professor" keeps showing the same gender or ethnicity, that is a sign of bias.

Compared 1 of 6