Why is "Made with AI" automatically added to my X posts? Can I remove it? Testing the conditions for the label

X has introduced [Made with AI] label to indicate that content was created using AI.

These labels are basically meant to be set by the author when posting, but some users have been reporting one after another that the label was added even though they did not set it themselves.

In this article, I will explain in detail why [Made with AI] labels are added automatically, whether labels that were added automatically can be removed, and under what conditions the labels are applied.

What is the [Made with AI] label?

X's [Made with AI] label is a label indicating that the content included in the post was created or edited using AI tools such as generative AI.

In recent years, with the spread of generative AI, issues such as low-quality mass-produced content known as so-called "AI slop" and fake images and fake videos have come to be seen as problems on X.

There are also users in fields such as art who prefer content created by humans over content generated by generative AI.

Under these circumstances, X implemented the [Made with AI] label at the same time as the "Paid partnership" label, which indicates that someone is promoting a product or service in exchange for compensation or incentives.

Although applying the label is basically optional, it is expected to serve as useful information when viewing content.

X adds AI labels and paid partnership labels to enhance content transparency

X has implemented a new feature "Content Disclosures" that lets you add labels such as "Paid partnership" and "Made with AI" to new or existing posts.

March 3, 2026

Why is the [Made with AI] label added automatically?

As explained in this article, the [Made with AI] label is basically designed to be set by the author from the posting screen when creating a post.

How to set the "Made with AI" label on X posts: explained separately for before and after posting

X (Twitter) has implemented a "Made with AI" label to indicate that a post (tweet) was created using AI. In this article, we clearly explain how to set the "Made with AI" label both before and after posting.

March 3, 2026

However, recently, some users have been reporting one after another that the [Made with AI] label was added to their posts even though they did not set it themselves.

This is caused by a mechanism that automatically detects whether images attached to posts were generated or edited with AI and automatically applies the label.

Regarding this feature, X's head of product, Nikita Bier, explained that they are trying to protect the global town square of the next era.

In other words, when the [Made with AI] label is added to a post even though you did not set it yourself, it is because X's system determined that the image in the post was generated or edited by AI.

Examining the conditions under which the label is added automatically

What kind of system does X use to determine whether an image was created using AI, and what standards does it use to apply the label?

When X's system determines that an image was created using AI, a toast notification saying [AI-generated content detected] appears, and internally this notification is named COMPOSE_VIEW_C2PA_AI_DETECTED_TOAST.

C2PA (Coalition for Content Provenance and Authenticity), which is included in this name, refers to what is known as Content Credentials, a technology for verifying the origin and editing history of content.

For images with embedded C2PA-compatible data, you can use dedicated tools to check things such as the camera used to take the image and the software used to edit it.

Images generated by OpenAI's ChatGPT, images generated by Google's Gemini, and images edited using Magic Eraser in Google Photo all contain C2PA metadata, making it possible to tell that they were generated or edited with those tools.

Judging from the name of the toast notification, it is possible that X determines whether an image was generated or edited with AI by reading this C2PA metadata.

I attached the following three AI-generated images to posts and tested under what conditions the label would be applied.

• Grok (does not support C2PA)
• Gemini (supports C2PA)
• ChatGPT (supports C2PA)

All of the images used in the test were judged to be AI with 99.9% confidence by Hive Moderation AI-Generated Content Detection, a well-known service for detecting AI-generated images.

When I uploaded these images to X, the images generated by Grok and Gemini were not labeled, while the image generated by ChatGPT was labeled.

However, on X there are also posts saying that images generated by Nano Banana (Gemini) were labeled automatically, and conversely, posts saying that a label was added even though no AI was used.

Therefore, while it is possible that X uses C2PA as one of the factors in determining whether an image was generated or edited with AI, it does not appear to rely on C2PA alone.

The specific mechanism and standards have not been made public, but based on the test results, it seems highly likely that X applies the label using its own system or criteria.

Can a label that was added automatically be removed?

In most cases, I think it is not a problem if the [Made with AI] label is automatically applied when you post an image that was actually generated or edited with AI.

However, if the [Made with AI] label is automatically applied even though no AI was used at all, there may be cases where you want to remove the label to avoid causing misunderstanding.

When you add the [Made with AI] label yourself, you can add or remove it from the menu not only before posting but also after posting.

When I checked the post containing the ChatGPT image that had been labeled automatically in the test mentioned above, the [Made with AI] label setting was grayed out and could not be changed.

In other words, whether the [Made with AI] label can be deleted or removed is as follows.

• If you set it yourself: it can be removed later
• If it was set automatically: it cannot be removed

Conclusion

Among some users on X, there have been many reports that [Made with AI] labels were added to posts automatically.

These labels are basically meant to be set by the author, but in some cases they appear to be applied automatically when X determines that the content includes material generated or edited with AI.

The specific mechanism and conditions have not been made public, but it is possible that metadata known as C2PA is used as one of the factors in that determination.

If the [Made with AI] label was set manually, it can be removed later, but if it was set automatically, it appears that it cannot be removed.