X’s comment downvote feature for Premium users is being tested as the platform moves to improve reply quality and reduce AI-generated spam.
The update introduces a thumbs-down option in the reply bar, allowing users to flag low-quality or misleading content. The feature is currently limited to paid subscribers. Unlike traditional systems, the downvote will not directly hide replies or impact like counts.
Instead, it collects user feedback to refine how replies are ranked. The goal is to personalise content while improving overall conversation quality. This approach marks a shift toward data-driven moderation.
When users tap the downvote icon, they can choose from several options. These include:
- Not Interested in This Post
- Incorrect or Misleading
- AI Generated
- Spam
- Report Post
This structure helps the platform gather detailed insights into user preferences and content issues. The rollout comes amid rising concerns over AI-generated spam and automated accounts.
By limiting the feature to Premium users, X aims to prevent coordinated misuse and bot-driven manipulation. Paid access adds a layer of accountability and reduces the risk of mass downvoting. Currently, Premium users represent a small portion of the platform’s overall audience.
X’s Head of Product, Nikita Bier, acknowledged weaknesses in the current reply system. He described it as lacking meaningful signals and often promoting low-quality content.
The new downvote feature is designed to address these gaps by introducing better feedback mechanisms. If successful, the feature could expand beyond Premium users or remain an exclusive tool.
