The 2026 terms of
Summary
- The 2026 terms redefine content to include prompts, outputs, and data “acquired or created” through X, with a worldwide license for any use, including AI training, without additional payment.
- New abuse clause focuses on AI ‘jailbreaking’, rapid injection and circumvention, while keeping damages at $15,000 per 1,000,000 messages within 24 hours.
- X retains Tarrant County, Texas as forum, imposes one- to two-year claim windows, class action claim waivers and a $100 liability limit, critics say cool investigation and litigation.
Social media platform
X and the crypto content
The revisions broaden the definition of user content to explicitly include “inputs, prompts, outputs” and information “obtained or created through the Services,” according to the draft terms. The current terms, dated November 15, 2024, will remain in effect until the 2026 version takes over.
Under the updated agreement, users grant Contents “for any purpose,” including analyzing it and training machine learning and AI models, according to the terms document. The agreement states that no compensation will be paid for this use and that access to the service constitutes “sufficient compensation.”
The 2024 terms framed user responsibility around “any content you provide,” without explicitly naming prompts and outcomes, a comparison of the documents shows.
The 2026 draft adds a prohibited conduct clause targeting attempts to circumvent AI. The terms define “abuse” and include attempts to circumvent platform controls through “jailbreaking,” “prompt engineering or injection,” language that is not included in the 2024 terms, according to the documents.
The updated agreement includes Europe-specific provisions related to the enforcement of content under European Union and United Kingdom law. The terms note that EU and UK law may require enforcement against content described as ‘harmful’ or ‘unsafe’, with examples including bullying or humiliating content, content about eating disorders and content about methods of self-harm or suicide. The 2026 Terms add UK-specific language describing how users can challenge enforcement actions under the UK Online Safety Act 2023.
X maintains restrictions on automated access and data collection, excluding crawling or scraping “in any form for any purpose” without prior written consent. The terms establish liquidated damages of $15,000 per 1,000,000 messages accessed, viewed or accessed in a 24-hour period when violations involve such a volume, according to the agreement. In the 2026 draft, the wording is adjusted so that it applies when a user causes or knowingly facilitates violations.
Dispute provisions require claims to be heard in federal or state courts located in Tarrant County, Texas. The 2026 text states that forum and choice of law provisions apply to “pending and future disputes” regardless of when the underlying conduct occurred. The updated terms establish a one-year term for federal claims and two years for state claims, replacing the prior version’s single one-year term.
The agreement includes a waiver that in many cases prohibits users from bringing claims as a class or in a representative proceeding, and limits X’s liability to $100 per covered dispute, according to its terms.
The Knight First Amendment Institute stated that X’s terms “will stifle independent research” and called the approach “a troubling move that the company must reverse.” The Center for Countering Digital Hate announced in November 2024 that it was leaving X pending a change in terms and criticized the Texas location requirement as a tactic to steer disputes to favorable courts. The Reuters Institute for the Study of Journalism described how lawsuits can have “a chilling effect” on critics, according to published statements.
The changes will take effect on January 15, 2026, according to X’s announcement.
