In Germany, several groups of specialists analyze the behavior of bots distributing false information in social networks. As a result of the elections, they came to unequivocal conclusions.

Bad news: fake news and accounts are still active. And yet the outcome of the last elections in the German parliament was decided by the voters themselves. This conclusion in the era of bots on the Internet, which disseminate inaccurate or simply false information, does not seem trivial to journalists and experts analyzing the activity in social networks.

Social Media Watchblog, Faktenfinder team of the first public television channel of Germany ARD, Botswatch company, Correctiv journalism bureau, Department of Political Data Analysis at the Technical University of Munich - only in Germany there are several groups of specialists who check the news on the one hand and analyze the influence of artificial intelligence on public opinion through the dissemination of false information, on the other.

Everyone can get a bot.

Under artificial intelligence, it is necessary to understand, first of all, bots - computer programs, the purpose of which is to distribute templates that are profitable to the creators, or words snatched from correspondence in social networks. These programs can create networks - botnets. They are manageable and create the appearance of a discussion, trying to hide that behind the alleged debate lies a set of computer algorithms. The exposure of bots takes time, while the news launched by them lives an independent life.

Anyone can create a bot: instructions on how to rewrite 15-20 lines of program code for their needs are in the public domain on the Internet. Or you can buy bots. Asking for the relevant request on the Internet, you will get a lot of suggestions. Prices on German websites start from two euros per bot in Instagram or Twitter. But these same services can be found on the Web and much cheaper: 10,000 imaginary users can be purchased for about 500 euros.

The bots have come out of hibernation There has been no significant activity of bots related to the elections in Germany this year, experts concluded both in Germany and abroad, for example, at Oxford University. But this does not mean that all fake accounts were asleep.

Recorded by Botswatch data show that on certain days, especially just before the day of voting on September 24, the activity of bots on the Twitter network was not inferior to the activity of live users. However, the tactics of fake users changed.

Not one big lie, but many small ones. In 2017, there was not one big story in Germany, such as a lie about "the girl Liza" or a rumor that everyone would remember, like the story that the Pope had allegedly blessed Trump on presidency. Instead, analysts have discovered "a stream of small rumors, inaccurate or false information messages," as described by Jutta Kramm of Correctiv Bureau. The overwhelming majority concerned refugees and migration - the topic where the right-wing populists in Germany from the Alternative for Germany specialty, who gained 12.6 percent of the vote in the Bundestag elections.

Channels for the distribution of messages on an emotional topic, like the refugee problem, has become the Facebook network, in addition to the not very popular in Germany Twitter. Because of its business model - the inclusion of advertising to users - Facebook is interested not so much in verifying the facts as in the longest stay of users on its page or in a mobile application, said DW author of Social Media Watchblog Martin Giesler (Martin Giesler).

This ultimately provokes the emergence of emotionally charged, but not necessarily accurate information, says Gisler. An example is the post of the AdG party after the terrorist attack in Barcelona with a bloody car wheel and accusations against Angela Merkel. According to the company's own data, before the election, its employees destroyed thousands of fake accounts that distributed unreliable messages.

Logic of "small lies" On the technical side, the meaning of the tactic of "small lies" is that small stories have little coverage. Distributed often in closed groups, they practically do not fall into the attention of local or federal media that could expose the misinformation. Without being exposed and staying on the Web, such messages reinforce the opinions of a certain group of people formed on the basis of inaccurate data, experts say..

On the content side, the bots used the political position that developed in Germany. "We had a situation two years ago when we accepted many refugees, partially unable to register and monitor them. In this situation, many people have a real fear, "- says Jutta Kramm.

Contrary to expectations.

Nobody knows for sure whether the recorded phenomenon was a "purposeful operation," warns Kramm. "I do not know for sure, but the feeling was created. Especially since consonant with this flow of lies published in social networks posts leaders of the party AdG, "- says journalist Correctiv.

There was a fixed activity of targeted bots, but the flow of "small lies" was what was least prepared in Germany, said DW Kramm: "Everybody was waiting for the bots to actively spread one big lie. The public was prepared. But I do not think that such a tactic would work here ".

The Russian trace? Too little evidence.

Contrary to expectations of interference from Russia, experts did not observe any significant activity from Moscow, except for the activities of bots on election day, as reported by experts from the Digital Forensic Research Lab, working at the Atlantic Council. But even this activity does not contradict the tactics of "small lies".

There are two explanations for weak activity from Russia:

the lack of interest in the destabilization of the leadership headed by Angela Merkel and the peculiarities of the German Information Society. "The majority of voters still believe the traditional media, so a massive propaganda campaign would have to be carried out only in social media, and this would not have the desired effect," says Kramm.

In Germany, according to the University of Oxford, social media as the main source of information is perceived by only 7 percent of the population, and more than two thirds claim that they do not use them to obtain information.