X deploys upgraded Grok AI to redirect stolen-content payouts to original creators

What happened

X is taking a tougher stance against creators who abuse its revenue-sharing program by reposting other people's content and soliciting engagement. According to Nikita Bier, X's newest version of the Grok AI model can detect duplicated content at three times the rate of its predecessor. The system is designed to catch not only reposted videos and images but also copies of viral text posts—including lightly edited versions where someone has added watermarks, intros, or other superficial changes to disguise stolen content as their own.

When the system identifies a stolen post, the monetized impressions will be redirected to the original uploader rather than the account that copied it. Bier said the company has already detected 1.5 million posts that were stolen. The crackdown also targets engagement bait—posts designed to game the platform's revenue-sharing program through manipulative tactics rather than genuine audience interest.

This move follows X's earlier efforts to encourage original content creation, including adding an improved video editor and recorder to the platform. The problem X is tackling is not unique: Instagram, Facebook, and Reddit have all implemented technical measures to detect reposted work and discourage content theft.

Why it matters

Creator monetization programs are only as credible as the systems that enforce them. If bad actors can simply repost viral content with minor edits and collect payouts, the economics tilt toward theft rather than originality—undermining the incentive for creators to produce genuine work on the platform. By deploying Grok to detect duplicates and redirect revenue to original creators, X is attempting to close a loophole that has plagued social platforms for years.

The choice to use Grok—xAI's own large language model with native access to real-time X data—is notable. It signals that X views content provenance and duplicate detection as problems its in-house AI is well-suited to solve, given its direct access to the platform's full post history and engagement signals. The threefold improvement in detection rate, if accurate, represents a meaningful step up, though independent verification of that figure is not yet available.

For creators, the change could mean that original work is more likely to be credited and monetized, while accounts built on reposting may see their revenue dry up. For advertisers and the broader platform ecosystem, it is a step toward cleaner content attribution.

What to watch

• Whether the redirected-payout system works reliably in practice, especially for edge cases like parody, commentary, or transformative edits that may be harder to distinguish from theft.
• How X handles disputes when a creator believes their original content was incorrectly flagged or when two accounts both claim originality.
• Whether the 1.5 million detected stolen posts lead to account-level consequences beyond revenue redirection, such as suspensions or reduced reach.
• How competing platforms respond—Instagram, Facebook, and Reddit already have detection tools, but X's approach of redirecting monetized impressions specifically is a distinct mechanism.