X Expands 'Content' Definition to Include AI Prompts and Outputs in 2026 Terms Update

Social media platform X has announced significant changes to its terms of service that will take effect on January 15, 2026. These changes broaden the company's definition of user "Content" and introduce provisions related to artificial intelligence, according to the updated agreement released by the firm.

Key Takeaways

Starting on January 15, 2026, X will redefine user "Content" to include all AI-related interactions, such as inputs and outputs.
Users will grant X extensive rights to use their Content for various purposes, including AI training, without compensation.
New anti-circumvention rules will prohibit attempts to bypass platform controls through methods like "jailbreaking."
The terms include specific regulations for EU and UK users regarding the enforcement of harmful content.
Restrictions on automated access and data scraping remain stringent, with significant penalties for violations.
Legal disputes must be resolved in Texas courts, with specific deadlines for filing claims.

X 2026 terms expand AI rights and anti-scraping penalties

X and the Crypto Content

The revised terms expand the definition of user Content to explicitly encompass "inputs, prompts, outputs," as well as any information “obtained or created through the Services.” The current terms, dated November 15, 2024, will remain effective until the 2026 version replaces them.

Under the new agreement, users will provide X with a global, royalty-free license to use, copy, reproduce, process, adapt, modify, publish, transmit, display, and distribute Content “for any purpose.” This includes analyzing such Content and training AI and machine learning models. Notably, users will not receive compensation for these uses, and simply accessing the service is regarded as “sufficient compensation.”

The 2024 terms focused primarily on user responsibility with respect to "any Content you provide" but did not specifically mention prompts and outputs, as can be seen by comparing the two documents.

The 2026 draft introduces a new clause that targets attempts to bypass platform controls, identifying actions such as “jailbreaking,” “prompt engineering,” or any other method that seeks to circumvent the policy. These terms aim to protect the integrity of X's service, contrasting with the language in the previous agreement.

Moreover, the updated terms contain provisions tailored to users in Europe, addressing content enforcement per EU and UK laws. The agreement stipulates that both European and British law may necessitate actions against content deemed “harmful” or “unsafe,” with specific examples including bullying, content related to eating disorders, and methods of self-harm or suicide. Additional UK-specific guidelines explain how users can contest enforcement actions under the UK Online Safety Act 2023.

Restrictions remain firm regarding automated access and data collection, explicitly prohibiting crawling or scraping “in any form, for any purpose” without prior written consent. The 2026 terms will impose liquidated damages of $15,000 for every 1,000,000 posts that are requested, viewed, or accessed within a 24-hour period if violations reach that scale.

Dispute resolution provisions dictate that claims must be addressed in either federal or state courts located in Tarrant County, Texas. The upcoming 2026 document states that these rules apply to “pending and future disputes,” independent of when the actions occurred. It also establishes deadlines for federal claims of one year and two years for state claims, updating the previous single one-year timeframe.

Additionally, the agreement contains a class-action waiver that restricts users from initiating claims as part of a group or in representative lawsuits in many circumstances. X's liability for covered disputes is capped at $100.

The Knight First Amendment Institute has criticized X’s new terms, suggesting they will hinder independent research and calling the changes a “disturbing move” that should be reconsidered. The Center for Countering Digital Hate also announced its decision to leave X ahead of the terms change, criticizing the Texas venue stipulation as a way to favor particular courts. Furthermore, the Reuters Institute for the Study of Journalism detailed how such lawsuits could impose a “chilling effect” on critics of the platform.

These changes will come into effect on January 15, 2026, as stated in X's announcement.