X has updated its terms of service to allow data published on the platform to be used to operate and improve its products. The terms also say collected data and publicly available information could be used to train machine learning and artificial intelligence systems, potentially giving xAI access to material posted by X users.
What X’s terms say about AI training
The change concerns data published on Twitter, now known as X, as well as publicly available information. Elon Musk has said that only public information will be used for AI training, not the content of private messages.
The distinction matters because the policy described in the source focuses on posts and other public material. It does not establish that every public post will necessarily be used in a model, or explain how individual posts would be selected.
X is also planning a job service that would collect information about users’ previous jobs and education, according to Bloomberg. That points to a possible expansion in the types of information the platform may handle, although the source does not say that this employment information would be used to train AI.
Why xAI could benefit
The company that could gain from the terms change is xAI, Musk’s AI startup. He unveiled the company in mid-April. It is described as separate from X Corp and says its aim is to use AI to “understand the true nature of the universe.”
Musk has also said xAI will create AI offerings for consumers and businesses, similar to OpenAI. Building large AI models calls for large amounts of useful data. X could provide a stream of public material, and the source says that data could be made available exclusively to xAI.
xAI says it plans to work closely with X, Tesla, and other companies. Its team includes people who previously worked at OpenAI, Google Research, Microsoft Research, and Deepmind. That planned collaboration makes the relationship between the social platform and the AI startup notable even though they are presented as separate companies.
Data is becoming part of X’s business strategy
Since Musk took over Twitter, he has made it harder for data to flow off the platform while also seeking to improve the quality of the material on it. The end of free access to Twitter’s API is one example. Access now costs between $100 and $5,000 per month depending on the package, while enterprise packages start at $42,000 per month.
Those prices change how outside organizations can access platform data. At the same time, the possibility of using public posts for AI training gives the material another potential role inside Musk’s companies. Data can serve both as content that attracts platform users and as an input to AI products.
X has also introduced incentives around posting directly on the platform. Musk rewards users who pay for visibility and publish content exclusively on X, and successful creators can receive a share of the advertising revenue they generate. He has promised journalists “more freedom to write and a higher income” if they publish directly on X.
What users and publishers should watch
The policy raises practical questions about how public posts may be used and who may benefit from that use. The source describes a right for X to use published data and says xAI could be a beneficiary, but it does not detail a process for opting out, identify the posts that will be included, or say how the resulting training data will be shared.
For publishers and creators, the incentives to post directly on X intersect with the platform’s potential value as an AI data source. Content may help creators reach audiences and earn advertising revenue, while also contributing to a pool of public material that could support machine learning systems.
The larger issue is whether X’s value lies partly in the data its users publish. The terms change gives that question a sharper edge: activity on the social network may have significance beyond what appears in a user’s feed, especially if it supports AI products developed by a closely linked company.