A new draft is now available for comments. This is the first release to contain a very early draft of the “checklist”. It’s based on the Model Openness Framework being developed by LF AI & Data/Generative AI Commons.
There were a few unresolved comments on draft 0.0.4 for which I opened separate threads. See
- Do we need to mention the level of autonomy of AI systems?
- Is the verb "Use" clear enough?
- Is the definition of "AI system" by the OECD too broad?
- Does the "Share" verb need clarification?
- Clarifications needed on the Out of scope section in draft 0.0.4
Changelog:
- Added the first draft of the checklist, based on the ongoing Track 1 exercise.
- Refactored lightly the Out of scope paragraph for clarity. More discussion needed.
- Left untouched other the preamble as more discussion is needed. More discussion needed.
- Left untouched the “Share” freedom. More discussion needed.