The selection process for the popular feature was shrouded in mystery until last week when Gizmodo reported former Facebook employees claimed news about conservative issues and people was routinely suppressed from Trending Topics.

While Facebook is investigating the allegations, both the company and Mark Zuckerberg have said that there was no evidence that Trending Topics was successfully manipulated. The Gizmodo report included claims of news curators injecting and blacklisting news.

To provide clarity to the situation, Justin Osofsky, Facebook vice president global operations, blogged on Thursday that algorithms are not the sole determinants for what ends up in Trending Topics. Humans are very involved too, and sometimes even RSS.

In Information about Trending Topics, Osofsky provided Trending Topics guidelines noting, “The guidelines demonstrate that we have a series of checks and balances in place to help surface the most important popular stories, regardless of where they fall on the ideological spectrum. Facebook does not allow or advise our reviewers to discriminate against sources of any political origin, period.”

Guideline highlights:

Algorithms first identify topics that have recently spiked in popularity on Facebook. “The Trending Topics algorithm also uses an external RSS website crawler to identify breaking events so that we can connect people to conversations on Facebook about newsworthy events as quickly as possible.”

The Trending Topics Team takes the topics identified by algorithms and then 1. Confirms that the subject is tied to a current event news event, 2. Writes a topic description that is corroborated by reporting from at least three of at a list of more than 1,000 media outlets, 3. Applies a category label to the topic and 4. Checks whether the topic is national or breaking news.

The Trending Topics list is then personalized for each via an algorithm based on a number of factors.

A Trending Topic, when clicked, goes to a search results page that includes related news and posts that have surfaced “algorithmically.” 

Osofsky also includes a FAQ that includes an explanation of injecting and when articles deemed “noise” are rejected, which is not blacklisting.