AI Glossary: Every Term Explained Clearly
A plain-language reference for AI terms. Each definition is honest, jargon-free, and links to deeper guides where relevant.
AGI (Artificial General Intelligence)
Hypothetical AI that can perform any intellectual task a human can. Distinct from narrow AI, which is good at specific tasks only. AGI does not exist as of 2026.
AI Hallucination
When an AI model states something confidently that is not true. The model produces text that matches patterns from its training data without any internal fact-checking.
Agentic AI
AI that takes actions on your behalf rather than just answering questions. It breaks a task into steps, decides what to do at each step, uses tools, and reports back. Synonym for autonomous AI.
Alignment
The research field studying how to ensure AI systems pursue the goals their designers intended, rather than misinterpreting goals in harmful ways. A sub-field of AI safety.
Anchor Text
The clickable text portion of a hyperlink. In SEO, descriptive anchor text helps search engines understand the linked page's topic.
Anonymous AI Chat
Using an AI chat tool without creating an account. Useful for evaluation, privacy, and quick questions. Most major tools require signup; a few (SentX, HuggingChat, DuckDuckGo AI Chat) allow anonymous access.
Autonomous AI
AI that takes actions on your behalf. Same concept as agentic AI. The core shift is from answering to doing — the AI plans, uses tools, and iterates.
Bleach (HTML sanitization)
A Python library for sanitizing HTML by allowing only a safe subset of tags and attributes. Used to prevent XSS in user-submitted or model-generated HTML.
Chatbot
A system that converses with you. Modern chatbots are powered by large language models. Distinct from an AI assistant, which takes actions beyond conversation.
ChatGPT Alternative
Any AI chat tool positioned as a substitute for ChatGPT. Common alternatives in 2026: Claude, Gemini, Perplexity, Copilot, SentX.
Context Window
The working memory of a single conversation — the maximum text a model can hold in one chat. Resets when the conversation ends. Distinct from cross-session memory.
Continual Learning
An AI's ability to keep learning after deployment without forgetting what it knew before. An unsolved research problem; modern assistants use memory layers instead.
Cross-Session Memory
An AI's ability to carry context — facts, preferences, project details — across separate conversations. The defining feature of memory-capable chat tools.
Deep Learning
A specific technique within machine learning that uses multi-layer neural networks. Modern language and image models are built on deep learning.
E-E-A-T
Google's quality framework: Experience, Expertise, Authoritativeness, Trustworthiness. Signals used to evaluate content quality, especially for YMYL topics.
Fine-Tuning
Training an existing AI model further on a smaller, specialized dataset to adapt it for a specific task or domain. Distinct from base training.
Generative AI
AI that produces new content — text, images, audio, video, code — rather than just classifying or analyzing existing data.
Hallucination
See AI Hallucination. When an AI states something confidently that is not true.
Hreflang
An HTML attribute that tells search engines which language and region a page targets. Essential for international SEO.
Image-to-Video
AI technique that animates a still image into a short video clip. More reliable than text-to-video because the model starts from a known frame.
In-Context Learning
An AI's ability to adapt within a single conversation based on examples you provide. Resets when the conversation ends — not the same as ongoing learning.
IndexNow
A protocol that lets websites notify search engines (Bing, Yandex) when content changes, for instant re-crawling instead of waiting for the next crawl cycle.
JSON-LD
JavaScript Object Notation for Linked Data. The recommended format for structured data on web pages — what search engines read to understand content.
Knowledge Panel
The information box Google shows for entities (people, companies, places) in search results. Driven by structured data + Wikipedia/Wikidata.
Large Language Model (LLM)
An AI model trained on enormous text data to predict the next word in a sequence. The foundation of modern chat assistants.
Machine Learning (ML)
A sub-field of AI where systems learn patterns from data rather than being explicitly programmed. Includes supervised, unsupervised, and reinforcement learning.
Meta Description
The HTML tag containing a page's description. Shown in search results under the title. Optimal length is 150-160 characters.
Memory (in AI)
An AI's ability to carry context across separate conversations. Distinct from context window (which is single-conversation only).
Narrow AI
AI that is good at a specific task only. Distinct from AGI. Every shipping AI product in 2026 is narrow AI.
OpenAI-Compatible API
An API endpoint that follows OpenAI's request/response format, allowing code written for ChatGPT to work with any compliant server.
Prompt Engineering
The practice of writing prompts that produce useful output from AI models. Despite the hype, mostly comes down to: be specific, provide context, specify format, iterate.
Reinforcement Learning
A machine learning technique where the system learns by trial and error, receiving rewards or penalties. Used in game-playing AI, robotics, and to fine-tune chat assistants.
Retrieval-Augmented Generation (RAG)
An architecture where an AI retrieves relevant documents from an external store before generating a response. Reduces hallucination by grounding the model in real data.
Rich Result
A search result that includes extra visual elements (ratings, FAQs, breadcrumbs) beyond the standard title + description. Driven by structured data.
Self-Learning AI
A broadly used term that can mean reinforcement learning, self-supervised learning, continual learning, or true self-improvement. In product marketing, usually refers to memory and in-context adaptation.
Self-Supervised Learning
A machine learning technique where the model learns from raw data without explicit labels, by predicting parts of the data from other parts. How modern LLMs are trained.
Sitemap
An XML file listing all URLs on a site for search engines to crawl. Lives at /sitemap.xml by convention.
Speakable
A schema.org property marking content suitable for voice assistants to read aloud. Eligibility for smart-speaker answers.
Structured Data
Machine-readable data embedded in web pages (typically JSON-LD) that helps search engines understand content. Powers rich results.
Text-to-Video
AI generation of a video clip from a text prompt alone. Distinct from image-to-video (which animates an existing still).
Token
The unit an AI model processes text in — roughly 4 characters or 0.75 words for English. Models have token limits per request.
Vector Database
A database optimized for storing and retrieving vectors (numerical representations of meaning). Used in RAG systems to find semantically similar content.
VideoObject
A schema.org type for marking up video content. Eligibility for video rich results in search.