OpenAI wants safeguards to remain effective when its models are accessed through partner services. In an October 21 statement, the company says hosting and reseller partners must maintain comparable protections, investigate possible violations and coordinate enforcement.
An updated version of its Usage Policies takes effect on November 20. Services that give third parties direct access to OpenAI models, including model routers and AI gateways, will need prior review and approval. The policy says an application is not covered solely because it integrates OpenAI models to power its features.
The update also explains restrictions on using model outputs to train other models and on consequential automated actions without human review. OpenAI presents these requirements as part of its safety approach. Its statement does not establish that every partner has fully implemented them.
OpenAI also says it is working with customers on SafetyIDs, identifiers intended to distinguish individual end users during investigations without revealing their identities to OpenAI. The company says this could target misuse more precisely and reduce disruption to other users. It does not announce completed deployment across all services or independently measured results.