Free ChatGPT-4o Images Might Soon Get Watermarks

OpenAI is currently testing a system of invisible markers for images created with ChatGPT-4o. According to reports, the new feature will only affect free users — their images might include subtle watermarks. This move, according to insiders, is a response to the surge in image generation following the rollout of ImageGen to all account tiers. The increase in demand has prompted OpenAI to look for ways to better manage content distribution.

ImageGen was previously only available to paid users, but with the expanded access, OpenAI is trying to strike a balance between openness and responsible use of the technology. These watermarks could become a tool to distinguish content generated by free users from that made using premium plans. That said, the company emphasizes that nothing is set in stone — the strategy may still evolve.

At the same time, OpenAI is working on an API for ImageGen, which would allow third-party apps to integrate image generation features. This opens the door for commercial uses, like design or educational tools. Experts point out that the current version of ImageGen outperforms rivals like Midjourney in terms of detail and handling complex prompts. If implemented, the watermark system could become part of a broader plan to manage server load and discourage unlicensed mass use of AI-generated content for commercial purposes.

For now, the markers are being tested privately. OpenAI says it will publish an official announcement once the changes are ready for public rollout.