OpenAI now requires prior company approval for platforms that give third parties direct access to its models. The revision to its usage policies, dated October 21, 2026, takes effect on November 20, 2026. Applications that merely use OpenAI models to power their own features are not covered, unlike services that resell or redistribute access to the models. The measure is part of a broader policy overhaul addressing model training using OpenAI outputs, protections for minors, high-stakes uses, and national security.
Source: Read the original article


