EU AI Act provider vs deployer: which one are you?

Published [Aug 6, 2026] · 3 min read

This post was made by a human.

Different transparency requirements, who does what.

In short, a deployer is whoever uses an AI system in their own business or work, while a provider is whoever built the AI system itself and put it on the market under their own name.

Some examples to clear it up for deployers:

• If you make an image on Midjourney, Grok etc., you are a deployer.
• If you ask Claude, ChatGPT, Gemini etc. to make the text for your post, you are a deployer.
• If ElevenLabs makes you an audio file, you are a deployer.

However, for providers the transparency obligations are a bit different:

• If you are OpenAI or Anthropic and make a new AI model for others to use, you are a provider.
• If you make an AI chatbot and sell that to others, you are a provider, while your users are deployers.

Needless to say, if you build an AI chatbot for your own application from the ground up, host it, and use it with your end users, you need to adhere to the transparency obligations for both providers and deployers.

The responsibilities also differ. For deployers, they are expected to track and label posts (if necessary). While providers should integrate the requirements into the solution.

The types of obligations therefore are as follows:

Providers are expected to build compliance into the tool itself (via a technical or a design requirement). Deployers are expected to disclose it at the point of use.

If you are a deployer and don’t want to bother with building your own compliant transparency log, just use our tool.