Updated introduction of Responsible AI in the 1.1 spec. #984
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅ |
benjelloun
reviewed
Dec 4, 2025
docs/croissant-spec-draft.md
Outdated
| 2. **Machine-readable RAI Data Documentation**: This specification proposes a machine-readable vocabulary for capturing and publishing existing Responsible AI (RAI) documentation solutions (such as [Data Cards](https://dl.acm.org/doi/pdf/10.1145/3531146.3533231)), thereby streamlining their publishing, sharing, discovery, and reuse. Further details are available in the [Croissant RAI specification](http://mlcommons.org/croissant/RAI/1.1). | ||
|
|
||
|
|
||
| We welcome additional extensions from the community to meet the needs particular and resposible AI aspects of specific data modalities (e.g. audio or video) and domains (e.g. geospatial, life sciences, cultural heritage). |
Contributor
There was a problem hiding this comment.
Typo: resposible -> responsible.
benjelloun
reviewed
Dec 4, 2025
docs/croissant-spec-draft.md
Outdated
| 1. **Data use and dissemination**: It provides a [set of mechanisms](#responsible-ai-and-governance) to enable the responsible use and dissemination of data. This is achieved by offering a machine-actionable representation of the data's provenance, lineage, and usage conditions at various levels of granularity. These mechanisms are built upon the integration of W3C standards (such as PROV-O and ODRL), ensuring compatibility with existing solutions. | ||
|
|
||
| 2. It records at a granular level how a dataset was created, processed and enriched throughout its lifecycle – this process is meant to be automated as much as possible by integrating Croissant with popular ML frameworks. By allowing the metadata to be loaded automatically, Croissant also enables developers to compute RAI metrics automatically and systematically, identifying potential data quality issues to be fixed. | ||
| 2. **Machine-readable RAI Data Documentation**: This specification proposes a machine-readable vocabulary for capturing and publishing existing Responsible AI (RAI) documentation solutions (such as [Data Cards](https://dl.acm.org/doi/pdf/10.1145/3531146.3533231)), thereby streamlining their publishing, sharing, discovery, and reuse. Further details are available in the [Croissant RAI specification](http://mlcommons.org/croissant/RAI/1.1). |
Contributor
There was a problem hiding this comment.
The RAI 1.1 spec doesn't exist yet... Do you want to link to RAI 1.0 for now?
docs/croissant-spec-draft.md
Outdated
| This is how Croissant helps address RAI: | ||
|
|
||
| 1. It proposes a machine-readable way to capture and publish metadata about ML datasets – this makes existing documentation solutions like [Data Cards](https://sites.research.google/datacardsplaybook/) easier to publish, share, discover, and reuse; | ||
| 1. **Data use and dissemination**: It provides a [set of mechanisms](#responsible-ai-and-governance) to enable the responsible use and dissemination of data. This is achieved by offering a machine-actionable representation of the data's provenance, lineage, and usage conditions at various levels of granularity. These mechanisms are built upon the integration of W3C standards (such as PROV-O and ODRL), ensuring compatibility with existing solutions. |
Contributor
There was a problem hiding this comment.
Let's just say provenance instead of provenance, lineage.
Contributor
There was a problem hiding this comment.
Also, can you link to the relevant sections of the spec?
docs/croissant-spec-draft.md
Outdated
| 1. **Data use and dissemination**: It provides a [set of mechanisms](#responsible-ai-and-governance) to enable the responsible use and dissemination of data. This is achieved by offering a machine-actionable representation of the data's provenance, lineage, and usage conditions at various levels of granularity. These mechanisms are built upon the integration of W3C standards (such as PROV-O and ODRL), ensuring compatibility with existing solutions. | ||
|
|
||
| 2. It records at a granular level how a dataset was created, processed and enriched throughout its lifecycle – this process is meant to be automated as much as possible by integrating Croissant with popular ML frameworks. By allowing the metadata to be loaded automatically, Croissant also enables developers to compute RAI metrics automatically and systematically, identifying potential data quality issues to be fixed. | ||
| 2. **Machine-readable RAI Data Documentation**: This specification proposes a machine-readable vocabulary for capturing and publishing existing Responsible AI (RAI) documentation solutions (such as [Data Cards](https://dl.acm.org/doi/pdf/10.1145/3531146.3533231)), thereby streamlining their publishing, sharing, discovery, and reuse. Further details are available in the [Croissant RAI specification](http://mlcommons.org/croissant/RAI/1.1). |
Contributor
There was a problem hiding this comment.
Hmmm.... How is this bullet point different from the one above? Aren't they both about creating machine readable RAI information?
Contributor
Author
|
I've uploaded a shortened version with only one bullet point and added graphical support from the PROV example to be consistent with other use cases. Please check it @benjelloun. |
benjelloun
approved these changes
Dec 5, 2025
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Since the first release, the approach of the Responsible AI (RAI) extension has evolved substantially.
This pull request updates the presentation of the RAI use case to maintain consistency with this revised approach.