Your AI Cheat Sheet: Breaking Down the Buzzwords

This resource will help you make sense of complex terms with simple, easy-to-understand definitions. Whether you're just starting your AI journey or brushing up on key concepts, this resource breaks down the jargon so you can navigate the world of artificial intelligence with confidence.

AI (Artificial Intelligence) is the science of making machines smart enough to do tasks that typically require human intelligence and time. Al has been running in the background of our lives for decades, suggesting words as we type, recommending songs in playlists, and recommending what we watch or buy based on our preferences.

Large Language Models (LLMs) are trained on massive amounts of written text and use machine learning to understand language patterns and relationships and how human language flows. LLMs can then predict which words should come next in a sentence to create extremely human-sounding responses. LLMs can translate languages, answer questions, summarize text, and even write stories, poems, and code.

Machine Learning is a key part of AI where machines learn from data to improve over time. Algorithms teach computers to recognize patterns, predict outcomes, and improve their over time by learning from data and feedback.

Reinforcement Learning means providing an AI model with positive or negative feedback helps it improve over time, sound more natural, and lead to better predictions and recommendations. For example, in ChatGPT, giving a response a thumbs up or thumbs down is your way of participating in its reinforcement learning process. Reinforcement learning from human feedback (RLHF) helps LLMs sound more conversational.

Generative AI means AI systems that create new content (text, images, music) based on the data it has been trained on. Unlike traditional AI, which may simply analyze or categorize data, generative AI can produce (or generate) novel, original outputs.

Multimodal AI is an advanced type of AI that can process and understand information from multiple sources or "modalities" like text, images, audio, and video all at once. ChatGPT and Google Gemini are multimodal models.

Prompts are the instructions you give to an AI model to guide its response (output). The quality and clarity of the prompt make a big difference in the accuracy and relevance of the AI's output.

Prompt Engineering means carefully phrasing a question (or prompt) to get the best possible response from an LLM.

Deepfakes are AI-generated images, videos, or audio files that convincingly mimic real people.

Hallucinations occur When AI systems generate information that sounds plausible but is actually false or nonsensical. This happens because LLMs, like ChatGPT, predict text based on patterns in their training data. They don't understand facts or context the way humans do, so they sometimes produce incorrect or made-up responses that sound convincing. Basically, the AI fills in the gaps with what seems like reasonable information, but it's just a guess, not a verified fact.

Ethical AI means developing and using AI systems responsibly and ethically, valuing human well-being, respecting fundamental human values and rights, and promoting society's greater good.

GPTs are pre-prompted, custom versions of ChatGPT that help with very specific tasks. You can use a pre-made GPT or easily build your own—no coding required! How to create a GPT