7 Total
1 High severity
5 Medium severity
1 Low severity
Summary

This is Hugging Face's documentation explaining how AI model creators should write 'model cards,' which are structured descriptions attached to every AI model published on the Hugging Face platform. The documentation describes required and recommended fields including the model's intended uses, limitations, potential biases, training data, evaluation results, and carbon emissions from training. If you download or use a model from Hugging Face, the model card is the primary document telling you what the model is designed for, what its known limitations are, and what license governs its use.

Technical / Legal Breakdown

This document is Hugging Face's Model Cards documentation page, a technical guidance resource governing how model creators should document AI models hosted on the Hugging Face Hub. The documentation states that model cards are 'essential' for providing information about models and describes a structured format including sections on model description, intended uses, biases, evaluation results, and training details. The document describes both mandatory and recommended metadata fields, including license declarations, language tags, dataset references, and carbon emissions disclosures, establishing documentation obligations for model publishers on the platform. The model card framework engages the EU AI Act's transparency and documentation requirements for AI systems, as well as broader responsible AI disclosure norms promoted by standards bodies; however, the document is technical guidance rather than a binding legal agreement, so the obligations it describes are platform-level publishing standards rather than legally enforceable terms. Compliance teams evaluating AI governance frameworks should note that model cards as described may constitute part of technical documentation required under the EU AI Act for high-risk AI systems, and the carbon emissions disclosure guidance may intersect with emerging ESG reporting requirements in multiple jurisdictions.

Institutional Analysis

Institutional analysis available with Professional

Regulatory exposure by statute, material risk assessment, vendor due diligence action items, and enforcement precedent. Available on Professional.

Start Professional free trial
High — 1 provision
Medium — 5 provisions
Low — 1 provision

Monitoring

Hugging Face has updated this document before.

Watcher includes same-day alerts, structured change summaries, and monitoring for up to 10 platforms.

Start Watcher free trial Or create a free account →

Professional Governance Intelligence

Need provision-level monitoring and regulatory mapping?

Professional includes governance timelines, compliance memos, audit-ready analysis, and full provision tracking.

Start Professional free trial

Cross-platform context

See how other platforms handle License Declaration Requirement and similar clauses.

Compare across platforms →

Mapped Governance Frameworks

California AB 2013 AI Training Data Transparency
US-CA
View official text ↗
DMCA
United States Federal
View official text ↗
DSA
European Union
View official text ↗
Archival ProvenanceSource & Archival Record
Last Captured May 12, 2026 06:32 UTC
Capture Method Automated scheduled archival capture
Document ID CA-D-000842
Version ID CA-V-002521
SHA-256 beb0e6cc72ab8a20d0c2cae62e1209b2ba7b7d6b14f4fd8aadf0ef4dd1234a7b
✓ Snapshot stored ✓ Text extracted ✓ Change verified ✓ Hash verified

Governance Monitoring

Monitor governance changes across the platforms you rely on.

Structured alerts for policy changes, governance events, and provision updates across 318+ platforms.

Create free account Compare plans