Company X Tests Tool That Reveals Hidden Post Limitations

Testing New Tool by Company X

According to НВ — Техно: On August 14 at 2:00 PM, Company X announced the start of testing a new tool that allows users to check whether the algorithm restricts the visibility of their posts. This feature is available on the 'Under the Hood' page in the app settings and is currently accessible only to a small group of testers. The tool provides the ability to upload account data and check the labels affecting the visibility of publications.

The service offers aggregated statistics for the past month in a JSON format file, rather than separately for each post. In the example presented by the company, two posts received NSFW labels, which led to their concealment from non-subscribed users and minors. Additionally, X disclosed part of the algorithm code that generates the 'For You' feed. The open code specifies labels for different types of content, such as:

• adult materials
• spam
• violence
• impersonation
• hate speech

Among the types of content, there is also a special label for materials related to crisis response. At the same time, no separate label for political content was found in the presented list, but there is a civic integrity label that may affect the visibility of a post, depending on the author's page. Before the publication of the new code, it was reviewed by external experts in recommendation systems, although X did not specify how the selection of experts was made or whether the results of their review would be disclosed.

Transparency of Social Media Algorithms

Interestingly, X had previously opened part of its recommendation algorithm code, but this did not provide a complete picture of the system's operation. The company also did not disclose data that could allow users to manipulate the algorithm. The code for the systems responsible for advertising and other functions unrelated to the 'For You' feed remains closed. In future updates, X may add new labels, including those related to brand safety.

The introduction of this tool is an important step towards the transparency of social media algorithms, which may increase user trust in the platform.

In the context of growing demands for accountability from technology companies, providing users with the ability to verify the visibility of their posts could become an important aspect in the fight against disinformation and content manipulation. However, despite the positive changes, the question remains regarding the full transparency of algorithms and their impact on user publications.

As transparency in algorithmic processes becomes increasingly vital, the recent move by Company X mirrors similar initiatives in the tech industry. For instance, Anthropic's introduction of an invisible code for AI outputs highlights the growing trend of making digital content more identifiable and accountable. This development not only enhances user trust but also raises questions about content moderation across platforms.