Concepts of information, misinformation and disinformation
How they differ?
It is essential to understand the related concepts of information, misinformation, disinformation and propaganda. The definition of information is clear by its very nature to the users. But what needs to be defined is the different forms it can take. We are more interested in the usage of social networks to spread specific kind of information to alter the behaviour or attitude of people. In the cyber space, manipulation of information so as to affect the semantic nature of information and the way in which it is interpreted by users is often called semantic attacks. Semantic attacks in social networks could be a result of propagation of information in various forms. This could take the shape of misinformation, disinformation or propaganda. The distinction between information, misinformation and disinformation is difficult to be made [3]. The three concepts are related to truth, and to arrive at a universal acceptance of a single truth is almost impossible.
The term information is defined by the Oxford dictionary as `facts provided or learned about something or someone’. The other forms of information are defined by Oxford dictionary as under:
-
Misinformation is false or inaccurate information, especially that which is deliberately intended to deceive.
-
Disinformation is false information that is intended to mislead, especially propaganda issued by a government organization to a rival power or the media.
-
Propaganda is defined as information, especially of a biased or misleading nature, used to promote a political cause or point of view.
The three definitions have small differences and the most important fact is they involve the propagation of false information with the intention and capability to mislead at least some of the recipients. The advent of social networks has made the speed of propagation of information faster, created large number of sources of information, produced huge amounts of information in short duration of time and with almost no accountability about the accuracy of data. The term `Big Data’ is often associated with the data in social networks. Finding credible information after sifting out the different forms of false information in online social networks has become a very challenging computational task. In this paper, we intend to use the basic tenets of cognitive psychology to devise efficient methods by which the task can be done. Our methodology involves detecting cues of deception in online social networks to segregate false or misleading information with the intention of developing an effective tool for evaluating the credibility of information received by a user based on the source of the message as well its general acceptability in the network.
Conceptual explanation of the distinguishing features
The concept of information, misinformation and disinformation have been differentiated with respect to five important features by Karlova et al. [2]. They are truth, accuracy, completeness, currency and deceptiveness. While all the three are informative in nature, only disinformation is deliberatively deceptive information. The authors have also given a social diffusion model of information, misinformation and disinformation as products of social processes illustrating the way they are formed and disseminated in social networks. The model suggests that people use cues to credibility and cues to deception to make judgements while participating in the information diffusion process.
Accuracy of the information is one of the important measures of quality of information. Honest mistake in the spread of inaccurate information is misinformation, whereas when the intention is to deceive the recipient, it is disinformation. In [4], authors have outlined the main features of disinformation.
-
Disinformation is often the product of a carefully planned and technically sophisticated deceit process.
-
Disinformation may not come directly from the source that intends to deceive.
-
Disinformation is often written or verbal communication to include doctored photographs, fake videos etc.
-
Disinformation could be distributed very widely or targeted at specific people or organizations.
-
The intended targets are often a person or a group of people.
In order to classify as disinformation, it is not necessary that the disinformation has to come directly from the source of disinformation [4]. In the chain of dissemination of information, most of the people could actually be transmitting misleading information (hence misinformation), though only one of the intermediaries may believe that the information is actually misleading (hence disinformation). This is especially true for social networks where the chain of propagation could be long and quite a few people involved in the process.
Social networks with its freedom of expression, lack of filtering mechanisms like reviewing and editing available in traditional publishing business coupled with high degree of lack of accountability have become an important media for spread of misinformation. Summarily, the propagation of different versions of information, viz misinformation, disinformation and propaganda involves the spread of false or inaccurate information through information diffusion process involving users of social networks where all the users may not be aware of the falsehood in the information. We have used the term misinformation to denote any type of false information spreading in social networks.
Misinformation
The acceptance of misinformation or misleading information by the people depends on their prior beliefs and opinions [5]. People believe things which support their prior thoughts without questioning them. The same is also supported by research in cognitive psychology [6]. The authors have brought out that preexisting political, religious or social views make people accept information without verification if it conforms to their beliefs. Countering such ideological and personal beliefs is indeed very difficult. Another important finding was that countering the misinformation may lead to amplifying the beliefs and reenforcing them.
Political astroturfing in the form of propagation of memes in Twitter was studied by the Truthy team [7],[8]. Investigating political election campaigns in US in the year 2010, the research group uncovered a number of accounts sending out duplicate messages and also retweeting messages from the same few accounts in a closely connected network. In another case, 10 different accounts were used to send out thousands of posts, many of them duplicates slightly altered to avoid detection as spam. With URL shorteners available, messages containing links could be altered to give different shortened links to the same source and hence escaping the spam filters.
Decision making out of ignorance is often based on heuristics and the level of confidence on the decision is also low, making correction easier. Such decisions are often correct and are generally not catastrophic. False beliefs based on misinformation are held strongly and often result in greater support for a cause. Such beliefs are also very contagious and the person makes efforts to spread them to others. The persistence of misinformation in the society is dangerous and require analysis for their prevention and early detection [6],[9].
Misinformation during an event as it unfolds like casuality figures in a natural calamity, are seldom accurate initially and the figures get updated or changed over a period of time. Such spread of misinformation is often considered benign though media is considered as one of the most important sources of misinformation. The other important sources of misinformation are governments and politicians, vested interests and rumours and works of fiction. Information asymmetry due to new media like social networks play a big role in the spread of misinformation. Social networks spread information without traditional filters like editing. The advent of Web 2.0 has resulted in greater citizen journalism resulting in increase in the speed of dissemination of information using multiple online social media like social networks, blogs, emails, photo and video sharing platforms, bulletin boards etc. The creation of cyber–ghettos has been discussed where the cyber space has become echo chambers and blogs and other social media primarily link to like minded sites propagating similar views than providing contrarian views. This leads to fractionation of the information landscape and consequent persistence of misinformation in large sections of the society for a long period of time. This often result in people holding on to their views on matters of pubic, political and even religious importance due to their misinformed world views and ideology.
In [10], authors have enumerated a number of possible instances of misinformation in the Internet. They include incomplete, out-of-date and biased information, pranks, contradictions, improperly translated data, software incompatibilities, unauthorized revisions, factual errors and scholarly misconduct. However, with the advent of Web 2.0 the list has grown many times and social media is described as one of the biggest sources of information including misinformation. Internet acts as a post modern Pandora’s box- releasing many different arguments for information which are not easily dismissible [11].
Countering the spread of misinformation
Misinformation is easily another version of information. Countering spread of misinformation is not an easy task. The simple technique of labelling the other side as wrong is ineffectual. Education of people against misinformation is necessary but not sufficient for combating misinformation. An analysis of the counter measures proposed and modeled in the literature against the spread of misinformation in OSNs are at times not in consonance with the effectiveness of the measures as suggested in studies of cognitive psychology. Theoretical framework for limiting the viral propagation of misinformation has been proposed in [12],[13]. The authors have proposed a model for identifying the most influential nodes whose decontamination with good information would prevent the spread of misinformation. The solution to the problem of limiting the spread of misinformation by starting a counter campaign using k influential nodes, called the eventual influence limitation problem has been proposed in [14]. The influence limitation problem has also been studied in the presence of missing information. In both the papers, the basic assumption is that when an infected node is presented with correct information, it would become decontaminated. Studies in psychology have proved that removing misinformation from infected persons is not easy [6]. The best solution to the spread of misinformation is early detection of misinformation and launch of directed and effective counter campaigns. In [15], the authors have proposed ranking based and optimization-based algorithms for identifying the top k most suspected sources of misinformation in a time bound manner.
-
The strategies proposed in [6] for effective counter measures include:
-
Providing credible alternative explanation to the misinformation.
-
Repeated retractions to reduce the effect of misinformation without repeating the misinformation.
-
Explicit warnings before mentioning the misinformation so as to prevent the misinformation from getting reinforced.
-
Counter measures be suitably biased towards affirmation of the world view of the receiver.
-
Simple and brief retractions which are cognitively more attractive than the corresponding misinformation.
Analysis of work done so far
We have analysed the cognitive process of adoption of information from studies in psychology. The difficulties associated with distinguishing between misinformation, disinformation and true information have been highlighted by most of them [2],[3],[6]. The cognitive factors which decide the credibility of messages and their consequent acceptance by users can be effectively modulated in OSNs as seen during US elections [7],[8]. The inherent beliefs of a user play a very important part in accepting news items and fractionation of cyber space is a consequence of this aspect of human mind. We explore different factors contributing towards deciding the credibility of news items in the next section.
Misinformation has been widely accepted in the society, it becomes extremely difficult to remove. This has been suitably demonstrated during July 2012, when mass exodus of thousands of people took place in India due to a sustained misinformation campaign by vested interests using social media and other telecommunication networks [16]. Preventing the spread of misinformation is a more effective method of combating misinformation, than its subsequent retraction after it has affected the population. Significant contributions towards successful debiasing of misinformation have been made in [6].
While studies in cognitive psychology are sufficient to understand the process of adoption of information by users, we would like to explore the process of diffusion of information. Process of diffusion is a group phenomenon, where we study the process of adoption by different users over a period of time. Patterns arising out of diffusion of information are better studied using algorithms from computer science. We study the process in detail using `Twitter’ in Section “Credibility analysis of Twitter”.
A generic framework for detection of spread of misinformation
While formulating a generic framework for detecting spread of misinformation, it is important to understand the cognitive decision making processes of individuals. A study of individual decision making process from a cognitive psychology point of view followed by a generic framework for detection of misinformation using the cues of deception is given in the following subsections.
Identifying cues to deception using cognitive psychology
The presence of misinformation in the society and real world social networks have been studied from psychological point of view extensively. An excellent review of the mechanisms by which misinformation is propagated and how effective corrective measures can be implemented based on cognitive psychology can be found in [6]. As per the authors, the spread of misinformation is a result of a cognitive process by the receivers based on their assessment of the truth value of information. Acceptance of information is more the norm than otherwise for most of the people. When people evaluate the truth value of any information they take into account four factors. The factors are characterised by asking four relevant questions. These questions are given below and illustrated in Figure 1, where we have summarised all relevant issues of misinformation.
-
1.
Consistency of message. Is the information compatible and consistent with the other things that you believe?
-
2.
Coherency of message. Is the information internally coherent without contradictions to form a plausible story?
-
3.
Credibility of source. Is the information from a credible source?
-
4.
General Acceptability. Do others believe this information?
Information is more likely to be accepted by people when it is consistent with other things that they believe is true. If the logical compatibility of a news item has been evaluated to be consistent with their inherent beliefs, the likelihood of acceptance of misinformation by the receiver increases and the probability of correcting the misinformation goes down. Preexisting beliefs play an important part in the acceptance of messages. Stories are easily accepted when the individual elements which make them up are coherent and without internal contradictions. Such stories are easier to process and easily processed stories are more readily believed. The familiarity with the sender of a message, and the sender’s perceived credibility and expertise ensure greater acceptance of the message. The acceptability of a news item increases if the persons are subjected to repeated exposure of the same item. Information is readily believed to be true if there is a perceived social consensus and hence general acceptability of the same. Corrections to the misinformation need not work all the time once misinformation is accepted by a receiver.
The rest of the paper is organised as follows. In Section “Research design and methodology” we give our research design and methodology where we explain the generic framework for the detection of misinformation in online social networks. As part of its implementation in `Twitter’, we carried out an analysis of the work done in estimating the credibility of information propagation in `Twitter’. In Section “Methods” we explain our methodology and algorithm for speedy detection of spread of misinformation in Twitter to aid a user to recognise misinformation and consequently prevent him from spreading it. In Section “Results and discussion” we show the results obtained using two different Twitter data sets. We outline our future work and conclude in Section “Conclusions”.