"Is it true that 'Russian bots have disappeared'? Joy spreads with the shutdown of Russian data centers"

The aggressive replies that usually flood in large numbers are somehow fewer. Could it be because the Russian data center has stopped, causing the opinion-manipulating bots to cease as well?—

Such a perspective is spreading on social media.

The trigger was an attack on the data center of the IT giant Yandex, known as the "Russian Google," and the subsequent observed sharp decline in automated communications. The graph's steep drop was clear and resonated with users' perceptions that the atmosphere on social media had changed.

However, there is an easily overlooked issue here.The "bots" appearing in internet statistics and the "bots" that pose as humans to post on social media do not refer to the same range of entities.

The reduction in automated communications and the decrease in disinformation or opinion manipulation—there are still stages that need to be confirmed between these.

What happened—Yandex's disruption and the decline in automated communications

According to an article published by The Independent on October 12, 2026, after Ukraine's attack on Yandex's data center, there was a noted decrease in communications from automated systems.

Cloudflare's data, introduced by the paper, showed a particularly significant drop in automated communications related to Yandex, nearing almost zero by Sunday morning. At the time of the article's publication, it had not returned to normal levels, and Yandex was reportedly working on service restoration.

What can be inferred from this is that at least Yandex's services and automated processes were significantly impacted.

On the other hand, on social media, this communication disruption was perceived as an event of greater significance. It was interpreted as Russian automated posting accounts stopping simultaneously, reducing aggressive posts in discussions on X and Facebook.

However, the original article points out precisely this leap in interpretation. A change in communication volume alone does not reveal what kind of posts, by whom, and how much they have decreased on social media.

Social media reactions—Voices saying "replies have decreased" and calm observations

According to The Independent, posts referred to as "bot tests" have begun appearing on X.

This involves posting opinions or claims that would typically receive numerous reactions from automated accounts to see how replies gather. Some users reported a significant reduction in certain types of replies, speculating that previous replies might have been automated posts.

Such reactions seem to reflect distrust towards the provocations and similar replies that occur routinely. However, this is an interpretation of the reaction's background and not a result of confirming the poster's identity.

Meanwhile, on Bluesky, an account called "antibot4navalny," which tracks bots and organized posting activities, challenged the explanation linking the attack on Yandex and Cloudflare's graph to the silence of Russian government-related bots and troll groups.

The reactions on X introduced here are based on the original article's report, while the reactions on Bluesky are based on verified public posts. This is not a survey targeting all social media users, nor is it an aggregation of Japanese users' reactions.

What can be confirmed is that while some perceive it as "quieted down," others question the interpretation of the data used for that explanation.

"Bots" are not just villains

When you hear "bot," you might imagine accounts that repeatedly post the same text or pretend to be human to spread political claims.

However, in a technical sense, bots are broader. They refer to systems where programs automatically process tasks instead of humans operating them one by one, and their uses can be beneficial or malicious.

A clear example is a search engine's "crawler," which automatically navigates web pages to gather information for search results. Such systems enable users to find the necessary pages.

Automated access is also used for site operation checks. On the other hand, automation is used for massive unauthorized access and information collection.

In other words,the mere figure of "reduced bot communications" cannot distinguish whether convenient automated processing has stopped or harmful activities have ceased.

If search-related crawling significantly decreases, the graph of automated communications can drop even if posts on social media remain unchanged. The volume of communication and the content of posts need to be verified separately.

Cloudflare's graph is not the "number of fake accounts" on social media

According to Cloudflare's official explanation, the Radar HTTP request data deals with communications received by websites using the company's services.

This is a useful observation for understanding internet trends, but it is not a statistic that investigates all accounts on X or Facebook to count fake accounts.

In the first place, the number of communications does not match the number of accounts. A single automated process can generate a large number of accesses. Conversely, if you want to grasp the activity of posting accounts, you need other information such as posting frequency, content, and coordination of multiple accounts.

Moreover, one should not read a graph about a company's network as a graph of all of Russia or the entire world's social media.

What should be noted is not just how sharply the line has dropped.

"What is being counted," "which network is targeted," and "compared to when." Without understanding these three, impressive graphs can easily spread misunderstandings.

"Bot tests" alone do not reveal the identity of the counterpart

So, does the experience of receiving fewer replies to one's posts mean nothing?

Not so. It can be a starting point for investigating the possibility of actual changes. However, it alone cannot determine the cause.

Suppose a post received 100 replies one day and only 10 on another. Possible factors include the posting time, interest in the topic, and the number of views. If the conditions for comparison differ, the number of replies will also change.

Even if replies decreased after the attack, it is necessary first to investigate "what kind of replies decreased." Then, it must be verified whether they were automated posts, who the operating entity is, and whether there is a connection with Yandex's disruption.

"Replies decreased," "previous replies were bots," "those bots were for Russian manipulation." These are separate claims.

Even if the initial change can be observed, the remaining two are not automatically proven.

In Japan, too, "a large number of replies = public opinion" is not always the case

This topic is not only related to overseas wars or IT companies. The same caution is necessary when reading Japanese social media.

For example, when a news item receives a large number of critical replies, it can appear as if it reflects the reaction of society as a whole. However, the people writing in the reply section are not randomly selected from the entire society.

Here, one should also be cautious of jumping to conclusions in the opposite direction. Just because there are disliked opinions or similar expressions, it cannot be concluded that they are foreign bots.

It is necessary to separate whether the content of the post is correct from who the poster is. Even real humans can spread false information, and automated posts can convey accurate information.

From the materials confirmed this time, it cannot be concluded that Russian-origin opinion manipulation has decreased in the Japanese language sphere. It is not advisable to bring overseas experiences directly into the explanation of the Japanese situation.

It cannot yet be definitively said "disappeared" or "zero impact"

Reading cautiously does not mean underestimating illicit activities on social media.

It is not necessary to deny the possibility that Yandex's disruption affected some illicit automated processes. However, that possibility and the explanation that specific impacts have been confirmed must be distinguished.

From the graphs and experiences introduced at this point, it cannot be said that "Russian opinion-manipulating bots have disappeared all at once." At the same time, that graph alone cannot prove "there was no impact at all."

Verification requires continuous analysis of posts on social media and investigations into the operation of fraudulent accounts.

What is being questioned in this commotion is not just the presence or absence of data. It is whether a convenient narrative is being added beyond the range that data can answer.

Even if there is a feeling that "the internet has become a bit quieter," it does not necessarily mean the reason has been clarified. Before sharing, confirm what the graph is counting. This small effort can prevent spreading another misunderstanding while intending to criticize false information.

Sources and References

1. The Independent (October 12, 2026)
   On the decline of automated communications after Yandex's disruption, "bot tests" on X, and the necessity to distinguish communication statistics from bot activities on social media.
   https://www.independent.co.uk/tech/bots-russia-x-facebook-meta-bluesky-twitter-b3065172.html

2. Cloudflare Official Document "Bots"
   Confirmation of the basic meaning of bots and the various uses of automated activities.
   https://developers.cloudflare.com/bots/concepts/bot/Cloudflare bot solutions docs

3. Cloudflare Radar Official Document "HTTP requests"
   On the observation targets of HTTP request statistics and the classification of human and automated communications. Confirmation that it is not a statistic that directly counts the number of social media accounts.
   https://developers.cloudflare.com/radar/investigate/http-requests/Cloudflare Radar docs

4. antibot4navalny's Bluesky post (October 11, 2026, UTC)
   Objection to the explanation treating Yandex's communication graph as evidence of the silence of Russian government-related social media bots and troll groups.
   https://bsky.app/profile/antibot4navalny.bsky.social/post/3mxmv7ka6xk27