Harvard Chan AI Glossary
In our AI courses, we often pause to explain the ideas behind AI and the language used to describe it. These posts grew out of those conversations. They’re for anyone who would like to understand AI better, with no technical background assumed.
1. Foundations of AI and Machine Learning
- What is artificial intelligence?
- What is AI literacy?
- What is artificial general intelligence?
- What is machine learning?
- What is supervised learning?
- What is unsupervised learning?
- What is reinforcement learning?
- What is classification?
- What is a feature in machine learning?
- What are neural networks and deep learning?
- How do predictive AI and generative AI differ?
- What is generative AI?
- What is computer vision?
2. Language models, architecture and inference
- What is an AI model?
- What is a large language model?
- What is a frontier model?
- How do open-source and open-weight models differ?
- How do local and hosted models differ?
- What is a transformer in AI?
- What is an attention mechanism?
- What is a mixture-of-experts model?
- What is a token in AI?
- What is AI inference?
- What does temperature mean in AI?
- What is reasoning effort?
- What is quantization?
3. Prompting, context and knowledge retrieval
- What are prompts and prompt engineering?
- What is a system prompt?
- What is zero-shot learning?
- What is few-shot prompting?
- What is chain-of-thought prompting?
- What are context and context windows in AI?
- What are context management and compaction?
- What does memory mean in an AI application?
- What are retrieval and retrieval-augmented generation (RAG)?
- What is semantic search?
- What is a vector database?
- What is a knowledge graph?
- Reading the names: model, harness, application and provider
4. AI agents, harnesses and operational infrastructure
- What is an API?
- What is an AI harness?
- What is a skill in an AI agent?
- What is state in software and AI?
- What is a subagent?
- What is a tool in an AI application?
- What is software architecture?
- What is JSON?
- What are permissions in an AI application?
- What is a plugin?
- What is a product requirements document?
- What is prompt injection?
- How do a repository, Git and GitHub differ?
5. Product design, engineering and workflows
- What are acceptance criteria?
- What is deployment?
- What is an experiment in product development?
- What is interaction design?
- What is iteration?
- What is a mental model?
- What is a minimum viable product?
- What is a pilot?
- What does production mean in software?
- What is a prototype?
- How do user experience and user interface differ?
- What is a user journey?
- What does validation mean in product development?
- What is a workflow?
6. Evaluation, evidence and human judgment
- What does accuracy mean in AI?
- What is algorithmic bias?
- What is automation bias?
- What is data leakage in machine learning?
- What is evaluation?
- What is exploratory data analysis?
- What is ground truth?
- What is an AI hallucination?
- What does human in the loop mean?
- What are product judgment and taste?
7. Practical guide to models, workspaces and providers
These posts distinguish model names from the applications and services around them. Examples are not rankings or endorsements. Capabilities and access depend on the model version and configuration.
- What are Claude models?
- What is DeepSeek?
- What are Gemini models?
- What is Gemma?
- What is GLM?
- What is GPT?
- What is Grok?
- What is Kimi?
- What is MiniMax?
Coding agents, harnesses and workspaces
- What is Antigravity?
- What is Claude Code?
- What is Codex?
- What is a command-line interface?
- What is Cursor?
- What is Goose?
- What is an integrated development environment?
- What is Lovable?
- How do Visual Studio Code and GitHub Copilot differ?
Applications, model access and choices
If you would like to find out more about AI and improve your leadership skills, consider the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.