AI Terms Glossary: Machine Learning, LLMs, and Deepfakes Explained

The core AI terms that show up constantly in the news these days, explained simply.

Artificial Intelligence (AI)

The broadest term for technology that replicates human learning, reasoning, and problem-solving in computers — machine learning and deep learning are its best-known sub-methods.

Machine Learning

A core AI method where, instead of manually programming every rule, a system learns patterns from large amounts of data and makes predictions on its own.

Deep Learning

A machine learning technique that stacks artificial neural networks — modeled loosely on the human brain — into many layers, dramatically improving image recognition and natural language processing.

Algorithm

The step-by-step procedure and rules for solving a problem; in AI, it refers specifically to the computational method used to learn from data and make predictions.

Training Data (Dataset)

The large collection of data used to teach an AI model to recognize patterns — the volume and quality of this data are considered the key factors determining AI performance.

Generative AI

AI that learns from existing data to create entirely new content — text, images, music, video — with ChatGPT and image-generation AI as leading examples.

Large Language Model (LLM)

A massive AI language model trained on enormous amounts of text data, capable of holding natural conversations and writing text much like a human.

Prompt

The instruction or question you enter to get the output you want from generative AI — the skill of crafting prompts carefully is known as "prompt engineering."

Transformer

A neural network architecture introduced in 2017 that uses an "attention" mechanism to capture relationships between words in a sentence simultaneously — it underlies nearly all of today's large language models.

Deepfake

Fake video or audio content that uses deep learning to convincingly synthesize a specific person's face or voice — its convenience comes with real concerns about potential misuse.

AI Hallucination

When an AI confidently presents false information as if it were fact — a key reason results from generative AI should always be independently verified.

AI Bias

When biases or imbalances present in the training data get reflected directly in an AI's judgments, producing results that are unfair or skewed against certain groups.

AI and Copyright

As generative AI trains on copyrighted material and produces similar output, how to protect original creators' rights has become an actively debated issue worldwide.

Artificial General Intelligence (AGI)

A hypothetical future AI that, unlike today's task-specific systems, could think and learn across every domain the way a human does.

Why Learning AI Terms Is Worth It

As generative AI like ChatGPT rapidly weaves itself into everyday life, AI terminology has stopped being specialist jargon and become common vocabulary in news and everyday conversation. Knowing the core terms makes it much easier to follow AI news and technology trends.

Frequently Asked Questions

Are "artificial intelligence" and "generative AI" the same thing?

No — artificial intelligence is a very broad concept, while generative AI is a more specific term for the subfield focused on creating new content.

Can I trust an AI-generated answer at face value?

No — as explained under "AI hallucination," generative AI can present false information as if it were fact, so important information should always be independently fact-checked.