Glossary · Rule layers and human rights
Provider usage policy (AI)
Also known as: Acceptable use policy (AI provider), AI provider usage policy, Prohibited use policy (AI)
A provider usage policy is the set of rules an AI provider publishes on how people may use its models and services — prohibited uses, high-risk use cases with extra requirements and the consequences of violations — which binds users through the provider's terms of service and is enforced by the provider, for example by restricting or closing accounts.
- AI regulation
- AI
In one sentence
A provider usage policy says how people may use an AI provider's models. It binds users by contract; it is not law and confers no conformity.
Example
Anthropic's Usage Policy prohibits, among other things, sustained and needless abusive or cruel behavior toward its models from November 12, 2026; a violation is a breach of the provider's rules, not of a law.
Explained in context
Context cards connect this term with others to answer one question. Also in British English and German.
- Provider rules, national law, EU law and human rights: who sets the rules for AIWhich texts govern what an AI system may say and do — and whom does each of them bind?UKDeutsch
- Claims about AI consciousness: four questions that get mergedWhich different questions hide behind a sentence such as "the AI is not conscious"?UKDeutsch
How it applies
- Addressee: The policy addresses people and organizations using the service — consumers, developers, businesses and, through them, the end users of integrated products. It does not address the model; that is the role of a model specification.
- Legal effect: It binds users by contract, through the terms of service that incorporate it, and within the limits of consumer protection, unfair-terms and platform law. It can be stricter than the law; it cannot set aside a statutory right or duty.
- Not the organization's own policy: An AI usage policy is a company's internal rule set for its employees' use of AI tools. A provider usage policy is the vendor's rule set for its customers.
- Examples (as of October 12, 2026): Anthropic's Usage Policy (version effective November 12, 2026), OpenAI's usage policies, Google's Generative AI Prohibited Use Policy (last modified December 17, 2024), Mistral AI's Usage Policy and DeepSeek's Terms of Use, whose disputes are governed by the law of the People's Republic of China.
- For documentation: Cite the policy with provider, title, version and date, and label it as a provider rule — never as a legal requirement or as evidence of conformity.