I think we’ve all seen the insane amount of “undress me” posts on X lately, using the new Grok AI feature, which allows anyone to ask Grok to manipulate any photo posted on a person’s timeline. But people have questioned if it was legal, and more importantly, they wanted to know how they could opt out. That isn’t going to be an option anymore.

If you post it on X, then X says they can use their AI to manipulate it.

Creators who publish work on X  are facing growing concern ahead of a major update to the service’s Terms of Service scheduled to take effect on January 15, 2026. Under the revised terms, users grant X Corp a broad, perpetual license to use nearly all content shared on the platform. This includes text posts, images, videos, artwork, AI prompts, and generated outputs. The license allows the company to copy, reproduce, adapt, analyze, distribute, and use that material for any purpose, including training and improving artificial intelligence and machine learning models. The updated language expands the definition of “user Content” to cover anything a user submits, inputs, creates, generates, posts, or displays on the service.

The license is worldwide, royalty-free, sublicensable, and irrevocable once granted. No separate opt-in is required. Continued use of the platform after January 15 is treated as consent.

According to the terms, users retain ownership of their work, but the rights granted to X are sweeping. The document explicitly states that content may be analyzed and used to train generative and non-generative AI systems, with no compensation paid to the creator. Deleting an account ends future submissions, but content already posted may remain covered under licenses granted prior to deactivation. The changes have drawn sharp reactions from artists, writers, photographers, and AI prompt engineers, many of whom rely on X as a primary platform for discovery and promotion. Visual artists, in particular, have expressed concern that portfolios built over years could be absorbed into future image generation models without credit or payment. A viral thread posted in late December by user @AubreanReverie helped accelerate attention to the issue. Replies to the post included accusations that the terms were unconscionable and calls for legal challenges. Some creators said they planned to stop posting original work, instead using X only to share links directing followers to personal websites or subscription platforms. Concerns have also been raised about AI prompts and generated outputs shared publicly on the platform. Some users believe even experimental prompt work could be swept into training datasets, reducing the incentive to share techniques or workflows openly. The update is not X’s first move to formalize AI training rights. Similar clauses appeared in earlier versions of the terms, but the 2026 update removes ambiguity by explicitly naming AI and machine learning uses. The timing follows recent platform changes that emphasize AI-powered tools, including expanded image editing and generation features introduced in late 2025. X has not announced an opt-out mechanism for creators who wish to continue using the platform while withholding content from AI training. As a result, some users are weighing whether to deactivate accounts entirely before the effective date.

What Users Can Do

Creators concerned about the new terms are being advised to review the full agreement carefully, download account archives if they plan to leave, and consider adjusting posting strategies. Common approaches include watermarking images, sharing lower-resolution files, or limiting posts to promotional announcements rather than original works. Legal experts note that users whose income depends heavily on intellectual property may want to consult an attorney specializing in copyright or digital media law. With the January 15 deadline approaching, the update has become a flashpoint in the ongoing debate over creator rights and the use of publicly shared content to fuel commercial AI systems.