3 QUESTIONS - Twitter's 'For you' impositions and perception operations

By Ersin Cahmutoglu

- The author works as a cyber security specialist at ADEO Cyber Security

ISTANBUL (AA) — In three questions, cyber security expert Ersin Cahmutoglu evaluates the perception operations of Twitter's personalized content.

• What do the new policies coming after January 2023 aim for?

Since acquiring Twitter, Elon Musk has practically rebuilt the company. Twitter's administrative organs have been changed in the course of this process, as have the critical technical units that serve as its backbone and employ qualified personnel. These changes also spanned key components like machine learning (ML), artificial intelligence (AI), and algorithms. The shift in policy since the first months of 2023, in particular, and the redesigning of algorithm systems have led to some ongoing debates today. Especially, millions of users around the world and in Türkiye have voiced the same complaint, questioning why accounts they do not follow and content that they are not interested in end up on their homepage.

Algorithms directly affect whether users and content on Twitter get featured or not. We have seen that many issues can be manipulated with even the smallest interventions. In particular, the information revealed by the "Twitter Files" disclosures that began in mid-December 2022 clearly demonstrated how algorithms were manipulated and what this caused. Since the steps for "redesign" taken by Musk in 2023, the content that users encounter on Twitter began to change.

As Twitter has said in its official statements, we see the content that the algorithms, which were redesigned after the new policies' adoption, want to show us. Twitter decides on this by itself and does not give us a choice in the matter. The main objective of these policies is to bring the most discussed and interacted topics and content on the platform to all users in a country.

The main debate here is on how algorithms are determined. According to experts, Twitter uses algorithms known as "black boxes." This means that the inner workings of the algorithms are kept secret from users. The platform features machine learning (ML) that sorts the content on our homepage according to "specific signals." We don't yet know what these signals are. Even Twitter engineers haven't entirely figured it out. Perhaps this is the main point that should be discussed. Because this situation points to an "operation."

• Is Twitter being used as a "tool of perception operation"?

The main basis of the numerous claims by users of a "political perception operation" on Twitter is that they constantly come across content by people they do not follow. There are many conspiracy theories and speculations on this subject. The main factor causing this irritating situation on Twitter is the algorithms. Only Twitter's executives can intervene and make changes in the algorithms in question.

When we examine the platform's own statements for more details, Twitter says it chooses the political tweets it shows us based on the "volume and health of conversations happening on Twitter ... We look at how much people are Tweeting, Retweeting, replying, and liking Tweets about that Topic."

From this statement, we can clearly deduce that when a Twitter user tweets positively or negatively about any political party, or comments or retweets other tweets, the algorithm sees this as a topic of interest for that person. So, positive or negative tweets about that political party are displayed on the person's home page. We see all of this content under the recently introduced "For you" tab.

The answer to how Twitter determines the content here is not just the algorithms. When we look at the institution's statement on this matter: "Your For you timeline displays a stream of Tweets from accounts you have chosen to follow on Twitter," explaining that tweets here are shown as a result of a technically determined ranking and signal algorithm. After this statement, Twitter says that these "For you" tweets are selected "using a variety of signals, including how popular it is and how people in your network are interacting with it ... This means you will sometimes see Tweets from accounts you don't follow.

As it can be understood from these statements, Twitter presents us with the content it wants through signal sorting and algorithms, in its own words, without asking its users and without getting their approval. In doing so, it has us follow certain topics automatically. In other words, it selects certain topics without our knowledge and brings them before us. At this point, the question what steps users can take comes to mind.

• What can users do against imposed content?

There are some settings that every user can use against these impositions by Twitter. We can disable topics and interests automatically determined for us by algorithms that are updated on the platform. However, it is worth noting that there is no guarantee we will be entirely rid of imposed content after these setting changes.

To apply these modifications, one must first access the "interests" and "topics" sections under "Settings" on Twitter. You may be unaware of the topics that you see there that have been chosen for you. This list is very long. You can disable any topic that Twitter automatically picks and marks, from politics to art, sports to technology. Recommended topics and interests in particular must be examined.

On the other hand, one can also mute some words in Twitter's settings. You will not see any content containing muted words. Most of those who do these settings report not getting most of the content they were uncomfortable with. To apply these settings, one can follow these steps one-by-one.

As can be seen, Twitter's new policies and the imposed content of the algorithms are of crucial significance in the users being manipulated and exposed to perception operations. At this point, it is necessary to emphasize the disinformation activities carried out on many social media platforms, especially Twitter.

As a result of a scandal that unfolded in recent weeks, we saw clearly that the politics in a country can be designed via Twitter with an army of "robot networks/avatars." It emerged that hundreds of thousands of robot networks, which are more advanced versions of bot accounts, carry out systematic activities as disinformation machines. In the cyber field, "advanced online influence operations," known as information operations (IO), will be one of the leading threats for many countries in the future and will bring along new discussions, especially in Türkiye as elections draw near.

* Opinions expressed in this article are the author's own and do not necessarily reflect the editorial policy of Anadolu.