X announced that its Terms of Service will change on January 15, 2026, expanding how the platform defines user “content” and adding contractual language related to the operation and protection of its AI systems.
The current Terms and Conditions, dated November 15, 2024, remain in effect until superseded by the 2026 version.
The key revision is that X now treats AI-era interactions as “content” for which users are responsible, along with posts and other materials.
How X’s updated terms redefine ownership and responsibility in the age of AI
Users are responsible for their content, including any “input, prompts, output,” and information “obtained or created through the Service,” and X cautions users to only provide, create, or produce what they are comfortable sharing.
The 2024 terms focused on “any content you provide,” without explicitly naming prompts or outputs. As such, Grok-style usage falls outside the scope of primary contract terminology.
This expanded definition will run alongside licenses that already grant X-sized reuse rights.
You grant a worldwide, royalty-free, sublicensable license to use, copy, reproduce, process, adapt, modify, publish, transmit, display and distribute the Content for “any purpose”, including analyzing the Content and training machine learning and AI models.
X also states that no compensation will be paid for these uses and that access to the services is “sufficient compensation.” This makes prompts and output language important for users who treat AI chat as separate from public posts.
The 2026 draft also includes specific prohibited acts aimed at circumventing AI.
“Exploitation” includes any attempt to bypass platform controls, such as through “jailbreaking,” “prompt engineering, or injection.” ”
Such expressions are not included in the similar list of misuses in the 2024 Terms. This provides a contract-based hook for X to cite when enforcing against attempts to breach AI feature safeguards, rather than relying solely on product rules or policy guidance.
European-specific languages change how documents describe content applications and user challenges.
The Overview and Content Regulations state that EU and UK law may require enforcement not only for content that is illegal, but also for content that is deemed “harmful” or “unsafe.”
Examples include bullying and humiliating content, eating disorder content, and content about methods of self-harm and suicide.
The 2026 Terms include additional UK-specific language explaining how users can challenge enforcement actions under the UK Online Safety Act 2023.
How X’s updated terms expand enforcement, data management, and user responsibilities
The updated terms maintain X’s limitations on automated access and data collection, including a liquidated damages schedule associated with mass browsing.
Crawling or scraping is prohibited “in any form and for any purpose” without prior written consent, and access is generally limited to “publicly available interfaces.”
These terms provide for a violation of $15,000 for every 1,000,000 posts requested, viewed, or accessed in a 24-hour period.
The 2026 draft adjusts the relevant language to apply in cases where a user induces or intentionally encourages a violation.
The dispute provision remains entrenched in Texas, but it has been modified in a narrow way that may extend some of the schedule of state law.
Disputes must proceed in the federal or state courts located in Tarrant County, Texas. The 2026 text adds that forum and choice of law provisions apply to “pending and future disputes,” regardless of when the underlying conduct occurred.
The 2024 terms more specifically mention the United States District Court for the Northern District of Texas as a federal venue option, along with Tarrant County state court.
The 2026 draft also splits the deadlines: one year for federal claims and two years for state claims. This replaces the single year clock in previous languages.
X also continues to limit how users can pursue claims and the amount they can recover if they win. This Agreement often contains class action waivers that prohibit you from bringing claims on a class or representative basis, and set X’s maximum liability at $100 per covered dispute.
These provisions have drawn criticism with widespread commentary on whether they reduce practical remedies even when users allege serious harm.
Critics warn that changes to X’s rules could stifle research and discourse
Public outcry over the changes has often centered on provisions that predate the 2026 draft and are still in place today, such as venue selection and the removal of penalties.
In a statement, the Knight First Amendment Institute said that X’s conditions “stifle independent research” and that this approach is a “disturbing move that companies should reverse.”
The Center to Combat Digital Hate announced in November 2024 that it was leaving Company X ahead of the changes, and criticized the Texas venue requirement as a tactic to steer disputes into more favorable courts.
The Reuters Institute for Journalism also notes how lawsuits can have a “chilling effect” on critics.
Concerns about AI training and licensing are being packaged as consumer hooks amid reports about users leaving the platform.
Provisions Current Terms of Use (November 15, 2024) Future Terms of Use (Effective January 15, 2026) What counts as “Content” User responsibilities centered on user-provided content AI workarounds that explicitly include “inputs, prompts, outputs” and information obtained or created through the Service No explicit “jailbreak” or prompt injection provisions “Including by “jailbreak” Prohibition on bypass attempts “Rapid engineering or injection” EU/UK Enforcement Framework No callout for UK Online Safety Act challenge process Adds “Harmful/Unsafe” example and remedy language of UK Online Safety Act 2023 US venue and claim window Northern District of Texas (Federal) or Tarrant County (State). One year term Tarrant County federal or state court. 1 year for federal claims and 2 years for state claims. Forum rules apply to pending and future disputes Scraping penalty of $15,000 per 1,000,000 posts requested, viewed, or accessed within 24 hours if related to a violation Same schedule, facilitation is limited to those that “lead or knowingly facilitate” users
With an effective date of January 15, 2026, X’s contract language will treat prompts and generated output as User Content under the Platform’s licensing and enforcement framework.
“Jailbreak” and immediate injection have also been added to the list of prohibited activities.
