Hugging Face requires model publishers to include structured metadata in their model cards, covering license type, language, and task type, and recommends additional fields for training data, evaluation results, and limitations.
This analysis describes what Hugging Face's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
The metadata fields establish the baseline disclosure a model publisher must provide, which directly affects how users understand what they are permitted to do with a model and what risks it may carry.
Interpretive note: The document describes these as platform-level publishing guidelines rather than enforceable legal terms, so the binding nature of compliance with specific fields depends on Hugging Face's separate terms of service.
This general introductory provision was removed as the guidelines shifted toward specific field requirements rather than general mandatory documentation principles.
View full change record →The mandatory metadata fields, including license declarations and intended use sections, determine what legal permissions and risk disclosures attach to each model a user downloads or integrates from the Hub.
How other platforms handle this
In assessing misuse, we consider factors such as evidence that a report was motivated by bias or hatred (e.g., based on protected characteristics such as race, sexual orientation, or gender identity) or other malicious intent.
We review account behavior and content that Members create, send, and publish in Mailchimp, including Campaigns and Websites.
We allow Content that names individuals in the highest positions in a company who have broad influence over the work environment, as long as the Content describes the individual's behavior or performance at work.
"Model cards are a form of documentation that provides information about a model's intended uses, how it was trained, its evaluation results, and other important details. Model cards are important for providing transparency about models and for helping users understand how to use them responsibly.Excerpt from Hugging Face's Model Card Guidelines
(1) REGULATORY LANDSCAPE: Mandatory metadata fields engage EU AI Act technical documentation requirements for AI systems, particularly for high-risk classifications.
Enforcement risk, jurisdiction flags, contract triggers, and due diligence action items.
Get the research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean.
The metadata fields establish the baseline disclosure a model publisher must provide, which directly affects how users understand what they are permitted to do with a model and what risks it may carry.
The mandatory metadata fields, including license declarations and intended use sections, determine what legal permissions and risk disclosures attach to each model a user downloads or integrates from the Hub.
ConductAtlas has identified this type of provision across 142 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Hugging Face.