X expands terms to treat AI prompts and outputs as user content

X is updating its Terms of Service to explicitly treat AI interactions as user content, a change that takes effect today, Jan. 15, 2026. The revision broadens the platform’s legal definition of “Content” to include “inputs, prompts, outputs” and any information “obtained or created through the Services,” bringing prompts and model outputs under the same rules that already govern posts, messages and uploads.

The company posted a summary of the changes to the X Privacy Center on Dec. 16, 2025, and the existing terms dated Nov. 15, 2024 remained in force until today. The new language makes clear that users are responsible for content they “provide, create, post, or otherwise utilize,” removing ambiguity about whether prompt text or results generated on the platform are private or excluded from moderation and reuse.

X’s update also tightens the license it claims over user content. Under the revised agreement, users grant X a worldwide, royalty-free, sublicensable license to use, copy, reproduce, process, adapt, modify, publish, transmit, display and distribute Content “for any purpose,” explicitly including analysis and training of machine-learning and AI models. The terms state no additional compensation will be paid for these uses and that access to the service constitutes “sufficient compensation.”

The company framed the changes as aligning contract language with evolving product features and legal obligations. "We’ve updated our Terms to better reflect that you are responsible for the content you post and create, including prompts, outputs, and/or information obtained when using X," the summary states, and it notes added prohibitions against interfering with the services.

Because AI inputs and outputs are now classified as Content, they are subject to the platform’s internal analysis, moderation and enforcement mechanisms. The terms add specific enforcement framing for jurisdictions with laws that require removal of harmful material, such as the United Kingdom’s Online Safety Act. In those cases X may make content unavailable only within the relevant country or geographic area while leaving it accessible elsewhere, a change that formalizes geo-targeted takedown and blocking practices.

The 2026 terms introduce explicit prohibitions aimed at protecting system integrity. Attempts to circumvent safeguards through “jailbreaking,” prompt engineering or prompt injection are forbidden, and automated access, crawling and scraping remain barred without prior written consent. The agreement sets liquidated damages at $15,000 per 1,000,000 posts requested, viewed or accessed in a 24-hour period when violations occur at that scale, and broadens the wording to cover users who induce or knowingly facilitate such breaches.

Other contractual updates include clarifications on governing law and dispute timelines for users outside the EU, EFTA states and the U.K., as well as notes that certain products may carry additional terms. The revised privacy section adds an “age assurance” measure allowing X to collect or share information to estimate or verify a user’s age where required by law.

The explicit inclusion of AI interactions in X’s content rules is likely to reshape how creators, prompt engineers and researchers use the platform and think about privacy. Key follow-ups for users and regulators include the final, operative clauses once the new terms are posted in full, guidance on treatment of interactions created before today, and technical details about how geo-restricted enforcement and prompt-related takedowns will be carried out.