X Detects 1.5 Million Stolen Posts in Latest Cycle
X has identified 1.5 million posts as stolen content during its most recent monitoring cycle, according to a statement from company executive Mark Bier. The platform is leveraging an upgraded version of its Grok AI model to detect duplicated material at three times the speed of its predecessor. This represents a significant escalation in X's efforts to protect original creators from content theft, a persistent issue across social media platforms. The company has not disclosed the exact time frame for this detection cycle, but the scale suggests a widespread problem. Bier emphasized that the new system can identify not just exact copies but also modified versions with watermarks or edited intros.
Over $1 Million in Payouts Redirected to Original Creators
X is redirecting more than $1 million in creator payouts back to the original authors of stolen content, Bier confirmed. Under the updated policy, any monetized impressions generated from duplicated posts—including those with watermarks, intros, or other superficial edits—will now be credited to the original uploader. This applies to both video and text content, with Bier citing the commonly stolen phrase 'Twitter is like the smoking section of the internet' as a frequent example. The financial redistribution marks a direct attempt to disincentivize content theft by removing its economic benefits. X's creator revenue-sharing program has been a key attraction for users, making this policy shift particularly impactful.
Three Strikes Policy for Engagement Bait Offenders
X is implementing a strict three-strikes policy for users who engage in solicitation of follows or interactions, commonly known as engagement bait. If a user is caught three times or more—for instance, by posting 'I'll follow everyone who replies'—their account will be permanently removed from the creator program. Additionally, the account will be forwarded to X's policy team for potential suspension. Bier has previously criticized engagement bait practices, even calling out top creator Mr. Beast for using financial incentives to drive viewership. The new rules apply to repeated or intentional attempts to circumvent the platform's policies, signaling a zero-tolerance approach to manipulative behavior.
Bot Suspension Accelerates to 208 Per Minute
X has ramped up its bot suspension efforts, now identifying and removing 208 bots per minute, according to Bier's April update. This aggressive pace reflects the platform's recognition that automated accounts are a primary vector for content theft and engagement manipulation. The company has been investing in AI-driven detection systems to complement its Grok-based content analysis. By targeting bots more quickly, X aims to reduce the volume of stolen content before it gains traction. The acceleration in bot suspension aligns with broader industry trends, as platforms like Meta and Reddit also deploy machine learning to combat spam and abuse. Bier noted that the rate is 'growing,' suggesting continued investment in this area.
Industry Context: Platforms Struggle with Content Theft
X's crackdown places it alongside other major platforms like Instagram, Facebook, and Reddit, which have implemented technical measures to deter content theft. Instagram and Facebook, for example, use tools that detect when users repost work without attribution. Reddit has also faced challenges with viral content being recirculated across its communities. X recently added an improved video editor and recorder to encourage creators to use native tools rather than stealing others' content. The platform's reliance on Grok AI for detection represents a unique approach, leveraging its own large language model to identify duplicates. However, critics note that enforcement remains reactive, and the effectiveness of these measures will depend on consistent application and user compliance.