AI Glossary

An AI system that can independently plan and carry out multi-step tasks, using tools and making decisions along the way rather than just responding to a single prompt.

The challenge of making sure AI systems actually do what humans want and intend, following our values and goals rather than finding harmful shortcuts.

When AI systems produce unfair or prejudiced results because of biases present in their training data or design.

AI tools that help programmers write, debug, and understand code by suggesting completions, answering questions, and even generating entire functions.

The study of moral questions surrounding AI development and use -- covering fairness, transparency, privacy, accountability, and the impact on society.

The broad field focused on ensuring AI systems do not cause unintended harm -- covering everything from preventing misuse to ensuring long-term safety of advanced AI.

Search engines enhanced by AI that understand the meaning behind your question and provide direct, synthesized answers instead of just a list of links.

Computer systems that can perform tasks that normally require human thinking, like understanding language, recognizing images, or making decisions.

A prompting technique where you ask the AI to show its reasoning step by step, which often leads to more accurate answers.

An AI program you can have a conversation with by typing messages, which responds in natural language just like texting with a person.

The maximum amount of text an AI model can consider at once -- including both your input and its response.

A powerful type of machine learning that uses layered neural networks to learn complex patterns from massive amounts of data.

AI-generated fake media -- typically video or audio -- that realistically depicts someone saying or doing something they never actually did.

A way of converting text, images, or other data into lists of numbers that capture their meaning, making it possible for computers to measure how similar things are.

A technique where you give an AI a few examples of what you want so it can understand the pattern and apply it to new inputs.

The process of further training an existing AI model on a specific, smaller dataset to make it better at a particular task or domain.

AI systems that can create new content -- text, images, music, video, or code -- rather than just analyzing existing data.

A family of large language models created by OpenAI that can generate human-like text, answer questions, write code, and much more.

When an AI model confidently states something that is factually incorrect or completely made up, as if it were true.

The broader field of AI creating images, encompassing text-to-image, image editing, style transfer, upscaling, and other visual creation techniques.

The process of actually running a trained AI model to get a response -- every time you send a message to ChatGPT, that is inference happening.

A massive AI system trained on enormous amounts of text that can understand and generate human language with impressive fluency.

A type of AI where computers learn patterns from data instead of being explicitly programmed with rules.

AI systems that can understand and work with multiple types of input -- such as text, images, audio, and video -- rather than just one.

The branch of AI focused on helping computers understand, interpret, and generate human language.

A computing system loosely inspired by the human brain, made up of layers of connected nodes that process information and learn patterns.

AI models whose code and weights are publicly released, allowing anyone to download, use, modify, and build upon them for free.

The millions or billions of adjustable numbers inside an AI model that determine how it processes information and generates outputs.

The skill of crafting clear, effective instructions for AI models to get the best possible results.

A technique that improves AI responses by first searching a knowledge base for relevant information, then using that information to generate a more accurate answer.

AI technology that converts spoken language into written text -- like a super-fast, highly accurate transcriptionist.

A setting that controls how creative or predictable an AI model's responses are -- lower means more focused, higher means more random and creative.

AI technology that generates images from written descriptions -- you type what you want to see, and the AI creates it.

AI technology that converts written text into natural-sounding spoken audio, producing voices that can sound remarkably human.

The small pieces of text that AI models read and generate -- like individual puzzle pieces that make up sentences.

A setting that controls AI randomness by limiting the model to choosing from only the most probable words that add up to a certain percentage.

The massive collection of information -- text, images, audio, or other data -- that an AI model learns from during its training process.

The neural network architecture behind nearly all modern AI language models, designed to process all words in a sentence simultaneously rather than one at a time.

A specialized database designed to store and quickly search through embeddings, enabling AI applications to find the most similar items in massive datasets.

When an AI model performs a task correctly without being given any examples -- relying entirely on its training knowledge and your instructions.