AI Dictionary

  • Tech that makes computers act smart, like humans do.

  • A software that does jobs for you on its own.

  • Making sure Al is used in ways that are right and fair to everyone.

  • Using Al to make tasks happen by themselves, without people doing them.

  • Step-by-step Instructions for solving a problem.

  • A way for different programs to talk to each other, often to use AI features

  • An Al even smarter than the smartest human ever.

  • A super-smart Al that can learn anything like a human

  • When an Al unfairly prefers some answers, often because the data it was trained on

  • A computer program that talks to people in text or voice.

  • Al that helps computers "see" and understand images or video.

  • A big collection of information that Al learns from.

  • Al that learns using brain-like structures called neural networks.

  • How easily people can understand why an Al made a certain decision

  • Training an Al a bit more on special data to make it better at specific tasks.

  • Al that can make new stuff, like pictures, writing, or music.

  • A famous type of Al that writes text like a human.

  • Special computer chips that help train and run big Al models faster.

  • Built-in checks to stop Al from making mistakes or causing harm

  • When Al makes up stuff that isn't true or isn't based on facts

  • When an Al uses what it learned to answer new questions.

  • A tag or answer given to data so Al knows what it is.

  • An Al model that understands and writes language, trained on lots of text.

  • A way for computers to leam things by looking at lots of

  • The final program that can do tasks after learning from data.

  • Rules for how Als share and use the Information given to them

  • Al that understands and works with human language.

  • Computer programs built a little like the human brain.

  • Al whose design is shared with everyone, so anyone can use or change it.

  • When Al leams the training data too well and can't handle new, different examples.

  • The text or question you give to an Al to get a response

  • Al learns by trying things and getting rewards for good actions.

  • Al that turns spoken words into written text.

  • An instance where an AI model (LLM) adapts responses to align with the user’s view, even if the view is not objectively true. A generally undesirable behaviour.

  • Tokens are the foundational, bite-sized units of text (words, parts of words, or single characters) that Large Language Models (LLMs) use to read, process, and generate language.

  • Using an Al trained for one job to help with a new, different job.

  • The process where Al learns from data and/or examples to get better at its job

  • A test to see if a computer can act so human that people can't tell the difference.

  • Al finds patterns in data that's not labelled

  • Specialised database designed to store, index, and query high-dimensional vector embeddings (numerical representations of unstructured data like text, images, and audio)

  • Al does a new task it wasn't directly taught just by understanding its description.