X (Formerly Twitter) to Allow Third-Party AI Training on User Data, Raising Privacy Concerns

X Is Giving Away Your Data to Train Third-Party AIs

Back when it was known as Twitter, the social media app had plenty of its own problems, but I'd wager the share of users who think X is a flourishing, functioning site in comparison is considerably small. The changes implemented under CEO Elon Musk have resulted in a platform that is less secure, more toxic, and, frankly, disrespectful of its users' privacy. You don't need to look very hard to see why X has no regard for your data: Last year, we found out that Circles, a feature explicitly designed for sharing posts to a private circle of followers, was just letting unapproved users see your posts for some reason. Then, back in July, X offered users the option to block the company's AI platform, Grok, from using their data for training -- suggesting the company may have been scraping user data without permission before that point. Now, it turns out X user data is again being used for AI training, but this time, for unnamed third-parties -- again, without prior consent. X seems to be selling your user data for third-party AI training As reported by TechCrunch, X updated its Privacy Policy on Wednesday, indicating the company will hand over user data to third-party collaborators for AI training purposes. This is the default, not an opt-in situation: If you don't like the sound of your posts and activity being used to train mysterious AI systems out in the world, you need to decidedly say so -- assuming you even know that's an option in the first place. While the rule change doesn't say whether X is benefitting financially from these third-party collaborations, that seems likely. After all, other companies have sold user data to other companies to train AI models, includiing both Reddit and Discord, each of which, like X, has an enormous amount of user-generated content to offer to top buyers. The specific portion of the Privacy Policy that calls out this activity is as follows: "Third-party collaborators. Depending on your settings, or if you decide to share your data, we may share or disclose your information with third parties. If you do not opt out, in some instances the recipients of the information may use it for their own independent purposes in addition to those stated in X's Privacy Policy, including, for example, to train their artificial intelligence models, whether generative or otherwise." What's potentially most egregious about this situation to me, aside from X selling user data in the first place, is that it's not obvious at all how to disable it. Even when you go to this setting's page, it says the following: "X always shares information with business partners as a way to run and improve its products. When enabled, this allows X to share additional information with those partners to help support running X's business, including making X's marketing activities on other sites and apps more relevant for you. Learn more." There's no explicit mention of selling your data for AI training. In fact, if you click that "Learn more" link, the page it leads to doesn't mention anything about AI training either. TechCrunch posits that X will update these descriptions on Nov. 15, when they officially update their Privacy Policy. Of course, that's speculation, so we'll have to wait and see. X doesn't care about you Either way, X clearly has no desire for the majority of its user base to opt-out of this data sharing, likely because it's in their financial interest to keep them in the dark. X is floundering when it comes to its finances, so the company needs to find ways to make money. Advertisers are fleeing, and most people don't want to pay for something they can get for free, so, sure, let's sell your data instead. TechCrunch also reports the new Privacy Policy removed language that defined the length of time the company would hold onto your data, in favor of language that says they will keep it for basically as long as your account is active, and as long as they see fit. In addition, you can't expect data your delete from X to be gone forever, as the policy says "public content can exist elsewhere even after it is removed from X." While the company uses examples like search engines and third-parties holding your data, you can imagine this also covers those third-parties who are training their AI models with your posts. If you've finally had enough, consider jumping ship to one of the many X alternatives out there. (And if you choose Bluesky, there's a good way to find all of your X followers there.)

Elon Musk's X is changing its privacy policy to allow third parties to train AI on your posts

On Wednesday, social network X (formerly Twitter) updated its Privacy Policy to indicate that it would allow third-party "collaborators" to train their AI models on X data, unless users opt out. While X owner Elon Musk trained xAI's Grok AI chatbot on X user data, leading to an investigation by the EU's lead privacy regulator, the company hadn't yet amended its policy to indicate its data may also be used by third parties. The addition to the policy implies that X, like Reddit and various media organizations, is looking into licensing data to AI companies as a potential new revenue stream. In Section 3 of the updated Privacy Policy titled "Sharing Information," X added a paragraph detailing how X user data can be used and how users could opt out. It reads: "Third-party collaborators. Depending on your settings, or if you decide to share your data, we may share or disclose your information with third parties. If you do not opt out, in some instances the recipients of the information may use it for their own independent purposes in addition to those stated in X's Privacy Policy, including, for example, to train their artificial intelligence models, whether generative or otherwise." The policy points to the settings page on X, but does not specifically indicate where users would go within the settings to toggle off data-sharing. Currently, the "Privacy and safety" section in settings lets users turn on or off data-sharing with xAI's Grok and with other "business partners," but the latter is described as those companies that X may work with to "run and improve its products," not other AI providers. That may be because the updated privacy policy won't become effective until Nov. 15th, at which point the opt-out option could be added. (We hope.) In addition, the company removed a paragraph that said it keeps user "profile information and content for the duration of your account," and that it keeps other "personally identifiable data we collect when you use our products and services for a maximum of 18 months." Instead, the new section explains that X will keep "different types of information for different periods of time, depending on how long we need to retain it in order to provide you with our products and services, to comply with our legal requirements and for safety and security reasons." As an example, it notes that usage information like the "content you post" and your interactions with others' content will be kept for "the duration of your account or until such content is removed." The policy also added a note reminding users that public content can exist elsewhere even after it's removed from X. This could potentially cover the data's ingestion by AI providers, as X adds, "search engines and other third parties may retain copies of your posts longer, based upon their own privacy policies, even after they are deleted or expire on X." Separately, X has added a new "Liquidated Damages" section to its updated Terms of Service that says any organization scraping its content will be liable for damages. Specifically, "for requesting, viewing, or accessing more than 1,000,000 posts (including reply posts, video posts, image posts, and any other posts) in any 24-hour period," X says the organization will be charged $15,000 USD per 1,000,000 posts.

X to allow third parties to train AI on your posts

It looks like even more AI is going to be trained on your shitposts on X. X announced back in July that it would use users posts to train X's AI chatbot Grok, but on Wednesday, the social network updated its Privacy Policy to allow third parties to train their AI models on X data, too. In the third section of the Privacy Policy, X writes that "depending on your settings, or if you decide to share your data, we may share or disclose your information with third parties." "If you do not opt out, in some instances the recipients of the information may use it for their own independent purposes in addition to those stated in X's Privacy Policy, including, for example, to train their artificial intelligence models, whether generative or otherwise," the policy reads. If you haven't already opted out of allowing X to train Grok on your posts, you can still do that. The new policy doesn't go into effect until November 15, 2024, and it doesn't look like there's any way to block X from allowing third parties to train their AI on your posts just yet. God willing, you'll be able to do this once it does go into effect.

X updates privacy policy to share user data with third parties

The updated privacy policy says that data may be used by third parties to train AI models. In a newly updated privacy policy, X has said it may share user data with third party "collaborators" unless users opt out. The policy includes a a paragraph on how X data can be used, stating that X may share or disclose a user's information with third parties. If users don't opt out, the recipients of X user data may use it for "their own independent purposes in addition to those stated in X's privacy policy," the policy now states, adding that it could be used to train their AI models. Additionally, the platform updated other sections about how long user data will be stored. The policy previously stated that "personally identifiable data" would be kept for a maximum of 18 months, however, it now states that the platform will keep "different types of information for different periods of time", depending on how long the platform needs it to provide users with services and to comply with legal requirements. How to opt out of sharing data with X To turn off data sharing by default in the X app, go to the 'Data sharing and personalisation' section under 'Privacy and safety' in 'Settings' and turn off 'Data sharing with business partners' and 'Grok'. X owner Elon Musk come under fire recently for using X user data to train the platform's AI chatbot Grok, leading to an investigation by Ireland's Data Protection Commission this August. Less than a month later, the DPC said the proceedings have been "struck out", due to X's agreement to suspend its processing of the personal data of the platform's EU and EEA users on a permanent basis. X has had a rocky time since being taken over by billionaire tech mogul Elon Musk. It was recently unbanned from Brazil after weeks of back and forth with the Brazilian judicial system over refusing to take down certain X accounts and failing to name a legal representative in the country - a legal obligation under Brazilian law. Don't miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic's digest of need-to-know sci-tech news.

X to allow third parties to train AIs with user data

Another raft of reasons to ponder your social media presence Elon Musk's social media mouthpiece X (formerly known as Twitter) has updated its Terms of Service and Privacy Policy to direct disputes to a federal court in Texas and allow third parties to train AIs on user posts. The updates will allow "third-party collaborators" access to user data, although it appears there will be some form of opt-out added. The specific paragraph reads: The policy does not clarify the setting for opting out of data slurping. There is an option for data sharing with business partners, but this seems more about advertising than anything else. No doubt there'll be something specific before long, which is opt out by default. The change to X's privacy policy comes into effect on November 15, 2024, and follows a growing trend of technology companies seeking to add revenue streams. For example, earlier this year, Reddit signed an AI training deal with Google after a paywall was thrown up around its API in 2023. Goodness knows X could do with an additional revenue stream after several significant brands opted not to use the platform for advertising following its transformation under the leadership of Elon Musk. Infamously, Musk told advertisers to go "f$ck themselves" in 2023, which was undoubtedly a contributing factor to the exodus of advertisers. That, and allegedly seeing their brands next to questionable content. And who could forget the time he threatened to sue Microsoft for allegedly using public Twitter data to train AI models, presumably referring to OpenAI's ChatGPT, although he ultimately did not do this. As we noted at the time, any model trained on public internet data might have ingested information via Twitter's API, which had historically been free until Musk started charging for API access. Also lurking in X's updates is a "Liquidated Damages" section in the Terms of Service, which appears to be aimed at anyone seeking to scrape content from the platform. Access more than 1,000,000 posts in a 24-hour period, and X will demand $15,000 per million posts. And if you have a dispute? It's off to Texas, specifically the US District Court for the northern district of Texas. While a move to Texas is not unexpected, considering X has moved its headquarters from San Francisco to the state, the selection of the northern district of Texas is curious, since X is now headquartered near Austin, Texas, which is covered by the western district of Texas US District Court. Some have theorized this is because the northern district is the seat of more conservative, business-friendly judges than the western district - criticizing it as "judge shopping." ®

X updates its privacy policy to allow third parties to train AI models with its data

The company is also imposing new penalties for scraping tweets. X is updating its privacy policy with new language that allows it to provide users' data to third-party "collaborators" in order to train AI models. The , which takes effect November 15, 2024, would seem to open the door to Reddit-like arrangements in which outside companies can pay to license data from X. The updated policy shared by X includes a new section titled "third-party collaborators." While the policy mentions the ability to opt out, it's not clear how users would actually do so. As TechCrunch , the policy points to users' settings menu, but there' doesn't appear to be an control for opting out of data sharing. The policy doesn't go into effect until next month, though, so there's still a chance that could change. X didn't respond to a request for comment. If X were to begin licensing its data to other companies, it could open up a significant new revenue stream for the social media company, which has seen from major advertisers. In addition to the privacy policy, X is also updating its with stricter penalties for entities that are caught "scraping" large numbers of tweets. In a section titled "liquidated damages" the company states anyone viewing or accessing more than a million posts a day will be subject to a penalty of $15,000. X owner Elon Musk has previously railed against "scraping." Last year, the company temporarily blocked people from viewing tweets while logged out, in a move Musk to fending off scrapers. He also moved X's API behind , which has drastically hindered researchers' ability to study what's happening on the platform. He's also used allegations of "scraping" to justify against organizations that have attempted to study hate speech and other issues on the platform.

Twitter/X will now allow third parties to train AI models with people's data -- and any disputes must be heard in a Trump-friendly court

Changes are on the way to the privacy policy at Twitter/X -- and they're pretty major. The Elon Musk-led social media company has alerted users that effective Nov. 15, user data can be used to train AI models for third-party "collaborators." That would seemingly extend beyond Musk's own Grok AI and let the company license data to outside companies, much like Reddit does. That could become a substantial revenue stream for the company, which has seen its advertising income tumble since Musk's takeover. The policy update does say users have the ability to opt out, but does not make it clear how to do so at this time. "Depending on your settings, or if you decide to share your data, we may share or disclose your information with third parties," the update reads. "If you do not opt out, in some instances the recipients of the information may use it for their own independent purposes in addition to those stated in X's Privacy Policy, including, for example, to train their artificial intelligence models, whether generative or otherwise." Should users have a problem with this (or anything else to do with Twitter/X) after the changes go into effect, they'll only have one venue to plead their case: the U.S. District Court for the Northern District of Texas or state courts located in Tarrant County, Texas. That's unusual for a few reasons. First, the company is not located in that county. It's headquartered 218 miles south in Bastrop, near Austin, Texas, following its move from San Francisco. Also, the judges in the U.S. District Court in that county were appointed by George W. Bush and Donald Trump. The county frequently sides with conservative parties in cases, whereas the District Twitter/X calls home has fewer Republican-appointed judges. Musk has become a super-donor to Trump and the Republican party, giving $75 million to his America PAC, a political action committee working to see that Trump is elected president in November. Twitter/X is also setting up protections to protect against data scrapers. Any user that is caught viewing or accessing over 1 million posts per day will be fined $15,000 per million, the company said.

X's New Rules: Blocked Posts Will No Longer Be Hidden, Your Tweets Will Train Grok AI

An upcoming revision to X's terms of service will make a recent plot twist at Elon Musk's social platform a black-and-white contractual reality: Your posts can officially serve as food for X's Grok chatbot and any other AI systems that X might create in the future. This practice has been documented at X since the company quietly updated a help file in July to note that your public posts (as well as your interactions with Grok itself) can be used to train X's AI models. That news, in turn, came a year after Musk revealed that his AI project would be trained on people's tweets. The new terms of service, which go into effect on Nov. 15, spell out the new policies. Under section 3, "Content on the Services," the terms require users to agree that X has the right to "analyze text and other information you provide and to otherwise provide, promote, and improve the Services, including, for example, for use with and training of our machine learning and artificial intelligence models, whether generative or another type." It's unclear whether that language means X will remove the current opt-out option that allows users to disable Grok's AI scraping. X did not immediately respond to a query emailed to its press-relations address, and a look over Musk's 100 most recent posts and replies did not reveal further details. That's not the only change to X's ToS that's raising questions. Under section 5, a new "Liquidated Damages" section gives the platform the right to fine users who use it too much. "Protecting our users' data and our system resources is important to us," the section says before declaring that "you will be jointly and severally liable to us for liquidated damages as follows for requesting, viewing, or accessing more than 1,000,000 posts (including reply posts, video posts, image posts, and any other posts) in any 24-hour period," at a rate of $15,000 for each million posts. If you keep doing that, this section adds, X reserves the right to enact "injunctive and/or other equitable relief, in addition to monetary damages." This policy is presumably meant to dissuade competitors from scraping X to train their own AI models, but some warn that it could also block researchers from using automated scraping tools for academic purposes. The revised terms of service also include new language that reflects X's recent move of its headquarters from San Francisco to the Austin suburb of Bastro. The company will soon require that all legal disputes be brought in "the U.S. District Court for the Northern District of Texas or state courts located in Tarrant County." Why? That federal court is based in Dallas (Austin lies within the Western District of Texas) and the county it resides in is next door to Dallas. But Musk and X have already gone out of their way to file lawsuits there against a now-defunct group of advertisers and the liberal advocacy group Media Matters for America -- both times choosing a small division of the Northern District in which a single particularly conservative judge, Reed O'Connor, hears cases. Lastly, the company also plans to change how blocking works on X. A post from the XEng account on Oct. 17 states that the block function will soon no longer hide your posts: "If your posts are set to public, accounts you have blocked will be able to view them, but they will not be able to engage (like, reply, repost, etc.)." A second post in that thread suggested that this move was a response to people using the block feature "to share and hide harmful or private information about those they've blocked." As of this publication, using the block option still yielded a dialog saying that blocked users "will not be able to follow you or view your posts." But with Musk having publicly said the old block function "makes no sense," it's likely only a matter of time before the change takes effect.