What is Post-Modern Computational Propaganda?

Blog Series about Detection of Computational Propaganda, part 1.

By Elnaz Babayeva and Sebastian Garcia

computational-propaganda-blog.jpg

In 1962, the respected philosopher Jacques Ellul said that we can not do targeted propaganda. Takes much too long” he said. “Much too difficult” he concluded.

modern propaganda will address itself at one and the same to the individual and to the masses. It can not separate the two elements. To address itself to individual, in his isolation, apart from the crowd is impossible

Ellul Jacques, “Propaganda: the formation of men’s attitudes “, 1973

Sixty years later, we are not only sure that we have the technological capabilities for targeted advertisement and propaganda, but we are also sure that it has been used for this purpose for a long time. The new targeted propaganda is possible thanks to the Internet, the mass collection of data and the new algorithms.

But before entering this new world of post-modern propaganda, we need to understand what propaganda is in its most basic and deep nature. We need to understand the political nature of this phenomenon so we can better detect it.

It is well known that Great Britain and the United States were the first countries to use mass media technologies for propaganda purposes, literally teaching Germany and the Soviet Union how to use new technology for that end. As citizens, we know that propaganda exists, but we don’t identify most of the current media manipulations as propaganda. We know that propaganda is dangerous and we don’t want to be influenced, but what exactly is propaganda?

What do people identify as propaganda?

Defining propaganda is difficult. This is because its definition changed and adapted with time, and different versions of propaganda are used by different countries and societies. The style of propaganda even depends on the cultural beliefs of those countries, since they may be based on different scientific bases.

There are even different types and stages of propaganda. Sociological propaganda sets the context, creates ambiguity and reduces prejudices. Direct propaganda comes later to modify opinions. Some propaganda is hidden and secret (black propaganda) and another is openly recognized (white propaganda). The problem defining propaganda relies on the fact that it is a broad umbrella concept covering a myriad of different activities. 

Although several definitions of propaganda have been proposed, most of them focus only in one aspect of it and therefore seem over-simplistic. For example, considering propaganda only as an information object (like a text in a tweet) fails to see propaganda as the conceptual idea of controlling and influencing a society. Even though propaganda uses information and news, it is more about the designs of the techniques of persuasion and control, than it is about the content of the text.

The danger of simplistic definitions is that they are not operational, meaning that they can not be directly used to identify or recognize propaganda from other types of communication. The usual definitions of what propaganda may be, such as appealing to emotions in the messages, may also match a lot of non-propaganda news.

As a comparison, imagine that you are trying to understand what the football sport is all about, all with its international competitions, complicated rules, international organizations, trades, politics, economic impact, scandals, and manipulations. And the definition you found of football is “22 players chasing a round ball in a court and trying to put it behind a goal line”. Although real, this definition does not help to understand the global phenomena. 


Instead of trying to define propaganda, we can try to see its high-level characteristics in a social space. Propaganda may be seen as a way of controlling social doctrine, as a way of maintaining social beliefs with multiple persuasion tactics. The three main characteristics that are usually recognized in propaganda are: 

  1. It is deliberately made to manipulate certain ideas.

  2. It presents only one view as the absolute truth.

  3. It uses psychological manipulation and convinces its audience that the propaganda’s aim and their own desires are one and the same.

These high-level characteristics are useful, but they are not operational definitions that can be used to detect propaganda. Therefore we will attempt an operational definition.

An Operational Definition of Propaganda

In order to improve the detection of propaganda with future algorithms, we can attempt to do an operational definition that help us understand how the information is being distributed and planned, and which are the patterns behind the creation of propaganda. 

We can identify the following characteristics, useful to identify propaganda from non-propaganda news. Propaganda:

  1. Must have an agenda. Sometimes it can be hidden.

  2. Can be based on true facts, but falsely interpreted.

  3. May be completely false.

  4. May use loaded language, instead of rational language.

  5. Must use repetition and several different media to be effective.

  6. Must be organized in a centralized and coordinated manner.

Moreover, propaganda may include other tools, such as fake news, rumours, disinformation, etc. Propaganda looks like a coordinated effort, that may use different techniques and ways of spreading.

In the last few years, the phenomenon of fake news grew in importance. But fake news is not propaganda. Fake news is a tool serving multiple purposes, and propaganda is only one of those.

If you still think that propaganda doesn’t influence you or doesn’t exist in your country, check the report done by the Computational Propaganda Research Project at the University of Oxford. This research shows that 70 countries are known to have used social manipulation; and in 26 countries, computational propaganda was used as a tool to suppress human rights, attack the opposition and to kill “not-right” ideas.

In the next blog we are going to talk about what is computational propaganda and how AI is helping to detect it.

Acknowledgment

This research was done as part of our ongoing collaboration with Avast Software.

 
avast-logo.jpg