How to Use AI Atlas

Find where a concept belongs

1. Search or browse for a concept.
2. Check its Level 1 parent.
3. Read the description.
4. Compare nearby sibling concepts.
5. Use the stable name in your lesson, article, brief, or discussion.

If a very specific term is not present yet, it may belong to future Level 3+ work.

Explain AI beyond ChatGPT

1. Open the Level 1 overview.
2. Show Generative AI as one branch among many.
3. Compare it with Machine Learning, Robotics and Embodied AI, Computer Vision and Perception, Search, Planning and Optimization, AI Safety, Alignment and Governance, and AGI and Future AI.
4. Ask: "What kind of AI are we actually talking about?"

Structure a lesson or curriculum

1. Use Level 1 as the broad scope map.
2. Pick a small number of branches relevant to the lesson.
3. Use selected Level 2 concepts as discussion topics.
4. Add your own examples, activities, and assessments.

AI Atlas gives structure, not a complete curriculum.

Check terminology before publishing

1. Identify the main AI claim or term.
2. Find the closest AI Atlas concept.
3. Check the parent area and description.
4. Avoid mixing fields, methods, architectures, applications, systems, products, and model families.

AI Atlas supports terminology checks, but it does not verify product claims or replace expert review.

Orient around AI Safety and Governance

1. Open the AI Safety, Alignment and Governance branch.
2. Compare subareas such as AI Alignment, Robustness, Interpretability, Evaluation, Bias and Fairness, Privacy and Security, Misuse Prevention, and Governance and Policy.
3. Use the branch as a discussion scaffold.

AI Atlas is not a legal compliance framework or risk register.

Teachers and educators

• Introduce AI as a broad field.
• Build a balanced lesson or module.
• Show that AI is more than ChatGPT.
• Create classroom discussion prompts.

Students and self-learners

• Build an initial mental model of the AI field.
• Find where a term belongs.
• Compare neighboring areas.
• Choose what to explore next.

AI Atlas is an orientation map, not a personalized learning path.

Researchers and taxonomy reviewers

• Inspect Level 1 and Level 2 boundaries.
• Identify gaps, overlaps, or unclear placements.
• Discuss concept type consistency.
• Prepare focused feedback.

Public feedback is welcome, but canonical taxonomy changes go through review.

Journalists, writers, and analysts

• Avoid vague "AI" language.
• Distinguish products from fields, methods, systems, and applications.
• Structure explainers and backgrounders.
• Clarify GenAI, Foundation Models, Agents, AGI, ML, and Safety terms.

Product and strategy teams

• Align team vocabulary.
• Tag internal AI initiatives by conceptual area.
• Avoid vendor-first or model-name-first strategy discussions.
• Ground AI claims and reduce AI-washing risk.

AI Atlas provides shared vocabulary, not a complete strategy process.

Policymakers and governance readers

• Orient around technical AI areas.
• Distinguish foundation models, generative AI, agentic systems, safety topics, and AGI discourse.
• Prepare neutral background explanations.

AI Atlas can clarify vocabulary, but it does not interpret laws or classify legal obligations.

AI safety and responsible AI readers

• Understand the substructure of AI Safety, Alignment and Governance.
• Separate AI Alignment, Robustness, Interpretability, Evaluation, Bias and Fairness, Privacy and Security, Misuse Prevention, and Governance and Policy.
• Keep responsible AI discussions anchored to specific concepts.

Visual material users

• Structure posters, slides, one-page maps, or workshop handouts.
• Keep materials tied to a public taxonomy version.
• Reduce reinvention of ad hoc AI landscape diagrams.

Downloadable assets should reference the AI Atlas version they are based on. See Download and Reuse AI Atlas.