Summary

In most cases, your AI vendor does not own your model outputs. Under the paid and business terms of OpenAI, Anthropic, Google, Microsoft, and AWS, you own the outputs or the vendor declines to claim them. Every vendor hedges that grant with the words “if any,” the ownership isn’t exclusive, and the consumer tiers treat your data differently, so “the output is yours” is the start of the answer and not the whole of it.

What do your AI vendor’s terms grant you?

On most paid or business tiers, an AI model’s usage terms give you ownership of the outputs. Each of the major vendors assigns you its rights in the output, states that you own it, or disclaims ownership outright, which is the answer founders look for in an AI IP ownership contract. OpenAI’s Business Terms assign the customer “all OpenAI’s right, title, and interest, if any, in and to Output,” Anthropic’s Commercial Terms say the customer “owns its Outputs,” and Google’s Gemini API, Azure OpenAI, and AWS Bedrock take the same approach for paying customers.

The vendors hedge for a reason. US copyright needs human authorship, and a raw output from a prompt often isn’t protectable, so in many cases nobody owns the copyright in the outputs of AI tools, as our guide to who owns AI-generated content explains in depth. You get permission to use the output without a copyright you can enforce against a copier. The ownership is also effectively non-exclusive. OpenAI’s terms note that output “may not be unique, and other users may receive similar content,” so the vendor stays free to generate the same or a similar result for another customer, including a competitor.

Owning the output doesn't mean your vendor defends you when it infringes. For more on this, see our write-up on AI Indemnification Clauses.

What does your AI vendor still control?

A hosted vendor doesn’t hand over the base model, and it keeps some control over how you use the outputs it grants you. You’re licensing the model rather than buying it, so the vendor owns the generative AI platform no matter how much you use it, and a model you fine-tune on OpenAI’s or Anthropic’s hosted infrastructure comes back as a vendor-hosted endpoint rather than portable weights.

The vendor can use what you submit on the consumer and free tiers, and may train on it, which is why those tiers don’t belong in a workflow that touches customer data. Most paid tiers also bar you from turning your outputs into a competing model, so you can use them but can’t build a rival on the vendor’s own service.

You’re likely fine here, but the details start to matter when you put your name on an IP promise in a customer contract or a financing. Serotonin Legal helps AI founders check what their vendors granted against what they’re agreeing to downstream, so reach out for a free consultation and we’ll give you a clear read on where you stand.

Reach out to set up a free consultation.

Serotonin Legal advises technology founders on corporate, regulatory, and transactional matters at the intersection of AI, blockchain, and fintech. This guide is for general informational purposes and does not constitute legal advice. No attorney-client relationship is formed by reading this material.