Tagoross logo

AI Openness:

Terminology and Definitions

A guide describing 21 key terms of the AI openness debate

Package 1 of 5

Available now

What do open source AI, open weight and open data really mean? This guide explains 21 key terms of the AI openness debate, one card each, grouped in five dimensions.

Free of charge. No registration needed.

PDF. Version 1.0 License: CC BY-SA 4.0.

AI Openness terminology guide cover

Same words, different meanings

A journalist, a regulator and a vendor can use the word "open" for the same AI model and mean three different things. That confusion involves risks and costs.

Before you use or invest in an AI model, you need to know which rights you really get, and which of them involve vendor lock-in, reducing your sovereignty.

This guide gives you the vocabulary to ask the right questions, in simple terms.

What you get in the guide

  • Key definitions. For each term: who defined it, what it says and what people often misread.
  • Five dimensions. Rights and licensing, scale and capability, measuring openness, public value, and drivers and failure modes. Terms that measure different things are kept apart.
  • A terminology map. All 21 terms on one page, to keep the big picture in view.
  • References and a glossary. Every source is linked, so you can check each claim.

More is coming

This guide is package 1 of 5. Together, the five packages give you a simple way to evaluate if an AI model is open or open source, and what that means for vendor lock-in and digital sovereignty. Each package answers one question:

  1. Terms and definitions. What does each term mean? (this guide)
  2. Timeline. How did the language of AI openness evolve, and why? (coming soon)
  3. How the terms relate. Where do people confuse or connect the terms? (coming soon)
  4. Classification matrix. Is a given AI model open, or open source? Where does it create vendor lock-in? (Tagoross Members only)
  5. Working with the terminology. How do I evaluate a model release myself, before my organisation depends on it? (Tagoross Members only)

Packages 2 and 3 will be published as open content (under a CC license), like this guide. You do not need to sign up to get them. We will announce each release on this page.

Want to see them before the release and help improve them? Join the review group with the form.

Join the review group

Each package is reviewed before its release. Enter your name and email, and check the box. It is free.

You will receive:

  • Pre-release versions of the resources, for your review.
  • A notification when each resource is published.

You may also be invited to join Tagoross.

We use your data only for what you choose in this form. Read our Privacy Policy.