Queen Essays

ScienceDirect

Available online at www.sciencedirect.com

Procedia Computer Science 141 (2018) 215–222

1877-0509 © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of EUSPN 2018. 10.1016/j.procs.2018.10.171

10.1016/j.procs.2018.10.171

© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the scientific committee of EUSPN 2018.

1877-0509

Available online at www.sciencedirect.com

Procedia Computer Science 00 (2018) 000–000 www.elsevier.com/locate/procedia

The 9th International Conference on Emerging Ubiquitous Systems and Pervasive Networks (EUSPN 2018)

Detecting Fake News in Social Media Networks Monther Aldwairi, Ali Alwahedi

College of Technological Innovation, Zayed University, Abu Dhabi 144534, UAE

Abstract

Fake news and hoaxes have been there since before the advent of the Internet. The widely accepted definition of Internet fake news is: fictitious articles deliberately fabricated to deceive readers”. Social media and news outlets publish fake news to increase readership or as part of psychological warfare. Ingeneral, the goal is profiting through clickbaits. Clickbaits lure users and entice curiosity with flashy headlines or designs to click links to increase advertisements revenues. This exposition analyzes the prevalence of fake news in light of the advances in communication made possible by the emergence of social networking sites. The purpose of the work is to come up with a solution that can be utilized by users to detect and filter out sites containing false and misleading information. We use simple and carefully selected features of the title and post to accurately identify fake posts. The experimental results show a 99.4% accuracy using logistic classifier.

© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Keywords: Fake news; clickbaits; social media; classification

1. INTRODUCTION

The idea of fake news is not a novel concept. Notably, the idea has been in existence even before the emergence of the Internet as publishers used false and misleading information to further their interests. Following the advent of the web, more and more consumers began forsaking the traditional media channels used to disseminate information for online platforms [11]. Not only does the latter alternative allow users to access a variety of publications in one sitting, but it is also more convenience and faster. The development, however, came with a redefined concept of fake news as content publishers began using what has come to be commonly referred to as a clickbait. Clickbaits are phrases that are designed to attract the attention of a user who, upon clicking on the link, is directed to a web page whose content is considerably below their expectations [24]. Many users find clickbaits to be an irritation, and the result is that most of such individuals only end up spending a very short time visiting such sites.

∗ Corresponding author. Tel.: +971-2-599-3238 ; fax: +971-2-599-3685. E-mail address: [email protected]

1877-0509© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Available online at www.sciencedirect.com

Procedia Computer Science 00 (2018) 000–000 www.elsevier.com/locate/procedia

The 9th International Conference on Emerging Ubiquitous Systems and Pervasive Networks (EUSPN 2018)

Detecting Fake News in Social Media Networks Monther Aldwairi, Ali Alwahedi

College of Technological Innovation, Zayed University, Abu Dhabi 144534, UAE

Abstract

Fake news and hoaxes have been there since before the advent of the Internet. The widely accepted definition of Internet fake news is: fictitious articles deliberately fabricated to deceive readers”. Social media and news outlets publish fake news to increase readership or as part of psychological warfare. Ingeneral, the goal is profiting through clickbaits. Clickbaits lure users and entice curiosity with flashy headlines or designs to click links to increase advertisements revenues. This exposition analyzes the prevalence of fake news in light of the advances in communication made possible by the emergence of social networking sites. The purpose of the work is to come up with a solution that can be utilized by users to detect and filter out sites containing false and misleading information. We use simple and carefully selected features of the title and post to accurately identify fake posts. The experimental results show a 99.4% accuracy using logistic classifier.

© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Keywords: Fake news; clickbaits; social media; classification

1. INTRODUCTION

The idea of fake news is not a novel concept. Notably, the idea has been in existence even before the emergence of the Internet as publishers used false and misleading information to further their interests. Following the advent of the web, more and more consumers began forsaking the traditional media channels used to disseminate information for online platforms [11]. Not only does the latter alternative allow users to access a variety of publications in one sitting, but it is also more convenience and faster. The development, however, came with a redefined concept of fake news as content publishers began using what has come to be commonly referred to as a clickbait. Clickbaits are phrases that are designed to attract the attention of a user who, upon clicking on the link, is directed to a web page whose content is considerably below their expectations [24]. Many users find clickbaits to be an irritation, and the result is that most of such individuals only end up spending a very short time visiting such sites.

∗ Corresponding author. Tel.: +971-2-599-3238 ; fax: +971-2-599-3685. E-mail address: [email protected]

1877-0509© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

216 Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 2 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000

For content publishers, however, more clicks translate into more revenues as the commercial aspect of using online advertisements is highly contingent on web traffic [12]. As such, despite the concerns that have been raised by readers about the use of clickbaits and the whole idea of publishing misleading information, there has been little effort on the part of content publishers to refrain from doing so. At best, tech companies such as Google, Facebook, and Twitter have attempted to address this particular concern. However, these efforts have hardly contributed towards solving the problem as the organizations have resorted to denying the individuals associated with such sites the revenue that they would have realized from the increased traffic. Users, on the other hand, continue to deal with sites containing false information and whose involvement tends to affect the reader’s ability to engage with actual news [4]. The reason behind the involvement of firms such as Facebook in the issue concerning fake news is because the emergence and subsequent development of social media platforms have served to exacerbate the problem [27]. In particular, most of the sites that contain such information also include a sharing option that implores users to disseminate the contents of the web page further. Social networking sites allow for efficient and fast sharing of material and; thus, users can share the misleading information within a short time. In the wake of the data breach of millions of accounts by Cambridge Analytica, Facebook and other giants vowed to do more to stop the spread of fake news [23].

1.1. Research Problem

The project is concerned with identifying a solution that could be used to detect and filter out sites containing fake news for purposes of helping users to avoid being lured by clickbaits. It is imperative that such solutions are identified as they will prove to be useful to both readers and tech companies involved in the issue.

1.2. Proposed Solution

The proposed solution to the issue concerned with fake news includes the use of a tool that can identify and remove fake sites from the results provided to a user by a search engine or a social media news feed. The tool can be downloaded by the user and, subsequently, be appended to the browser or application used to receive news feeds. Once operational, the tool will use various techniques including those related to the syntactic features of a link to determine whether the same should be included as part of the search results.

2. LITERATURE REVIEW

A look at contemporary scholarly work shows that the issue of fake news has been a major concern amongst scholars from various backgrounds. For instance, some authors have observed that fake news is no longer a preserve of the marketing and public relations departments [21]. In the stead, the problem is increasingly being regarded as part of the responsibilities associated with the information technology (IT) department. Traditionally, it was believed that the two departments mentioned above were the ones to deal with any implications arising from the dissemination of misleading news related to an organization. However, current research indicates that fake news is considered to be a threat to information security. The involvement of the IT department, therefore, is premised on the idea that it would help avert the various risks associated with the problem. Similarly, other authors have noted that the participation of IT professionals in resolving matters concerning fake news is paramount considering the demands of the contemporary corporate environment [7]. Rather than as it was the case a few years ago when perpetrators of such gimmicks were motivated by just attracting web traffic, the practice has evolved into a matter that includes the involvement of hackers. Specifically, some content publishers have resorted to including material that contains malicious code as part of the content provided on their web pages, leading those who visit such sites to click the links and download the malware without their knowledge. Such developments, according to the scholars, have exposed modern companies to further risk of cyber intrusion as the perpetrators of the fake news tend to target employees of certain organizations with the aim of exploiting the latter’s curiosity [2].

Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 217 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000 3

It is also apparent that aside from the risk of having malware introduced into their information management systems, modern firms also have to deal with the challenge of having their employees manipulated into giving out their credentials. Some scholars have posited that there is a group of content publishers that is increasingly using clickbaits as a technique to facilitate their phishing objectives [17]. Once an individual, who also happens to be an employee of the target firm, clicks on the link and accesses the web page’s contents, he or she is led into providing sensitive information, albeit in an indirect manner. The user may, for instance, be tricked into believing that they are helping to disseminate the news further when, in the actual sense, they are providing the perpetrators with access to their emails [19]. Data integrity has also been singled out as being one the information security implications associated with fake news [18]. In the current business world, data is increasingly being considered as being a valuable asset and, as such, it is imperative that companies put in place all the necessary measures that would help secure sensitive information from being accessed by unauthorized persons. However, the prevalence of content publishers keen on using fake news serves to negate such efforts. It is against this background that organizations are investing more resources to facilitate the invention and formulation of more effective solutions to be used in countering the ramifications that arise from using clickbaits to attract users into providing their information. Nonetheless, employees still continue to visit such sites even after being discouraged from doing so and, thereby, placing their firms at risk of cyber-attacks [6]. On the other hand, some scholars have argued that fake news can sometimes result in positive implications. For instance, there have been cases whereby companies listed in the stock market have experienced an increase in the price of their shares as a result of fake news [13]. As more and more users share the link to the site containing information that is seemingly related to an organization, prospective investors gain interest in the firms operations and, consequently, its share price increases considerably. Such changes, however, are bound to result in worse consequences as a majority of the individuals who buy the shares based on the misinformation end up being disappointed. In the same vein, other authors have noted that fake news can help further the marketing objectives of an enterprise. For example, when the information provided in the web pages associated with such news is one that favors the products furnished by a company, more consumers develop an interest in the same despite the fact that the contents of the web page are far from the truth [15]. Regardless, such an organization ends up reaching out to a wider pool of prospective clients in spite of the fact that the fake news was not part of its marketing campaigns. The scholars posit that the concept of fake news is not bad in its entirety as it can contribute positively toward the growth of an enterprise. However, this tendency has its limits and cannot be relied upon by businesses as its opposite would have extensive and adverse ramifications [8]. When the contents of the web page contain misleading information that portrays a company in a negative light, such a firm is bound to experience a drop in its performance irrespective of the fact that the news disseminated to its prospective customers was false. It is also apparent that the idea of using clickbaits to lure non-suspecting users to visit web pages has played a significant role in shaping opinions within other contexts aside from that which involved the business environment. For instance, the events leading to the 2016 presidential elections of the United States were characterized by the widespread dissemination of fake news through social media platforms [9]. Claims of celebrated personalities endorsing certain candidates were, for example, part of the information that was being shared by the users after visiting sites that informed them of the same. Later on, the users would realize that the assertions had been false. By then, the intended impact would have already occurred, and it is argued that such occurrences might have played a contributive role in determining the course of the elections [1]. Finally, the contemporary literature indicates that there have been ethical concerns about the whole concept of fake news especially regarding the involvement of individuals who have a background in journalism. For instance, some scholars have argued that using clickbaits is a demonstration of a disregard for the ethics associated with the media profession [16]. Journalists are expected to furnish readers with information whose veracity and accuracy have been determined to the last detail. However, the idea of fake news is completely at variance with these requirements. When professionals engage in activities that are intended to misguide their readers for the sake of increasing web traffic and online ad revenues, it raises a concern as to whether such people are keen on complying with the code of conduct associated with their career.

218 Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 4 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000

Despite fake news detection in social media getting attention fairly recently, there has been a flux of research and publications on the issue. Before talking about machine learning for fake news detection we must address the dataset issue. William Yang Wang [26] in his paper ”Liar, Liar Pants on Fire”, provided a publicly available dataset and so did many of the previous researchers. Additionally, the first Fake News Challenge Stage-1 (FNC-1) was held in June of 2017 and featured many novel solutions using various artificial intelligence technologies [? ]. Natural Language Processing (NLP) techniques have been used for news outlet stance detection to facilitate fake news detection on certain issues [20]. Riedel et al. and other FNC-1 winning teams achieved close to 82% accuracy in the stance detection stage. Once this competition and all stages of fake news detection are concluded, we believe great and commercial solutions will emerge. FNC-1 have made the datasets available publicly and we’re getting closer to having standard benchmarks to compare all the newly proposed techniques. For a more comprehensive survey of work on fake news detection, the reader is referred to Kai Shu et al. [22]. In this effort we try to focus on a lightweight detection system for clickbaits based on high-level feature title features.

3. Proposed Solution

The proposed solution involves the use of a tool that is designed with the specific aim of detecting and eliminating web pages that contain misinformation intended to mislead readers. For purposes of attaining this goal, the approach will utilize some factors as a guide to making the decision as to whether to categorize a web page as fake news. The user will, however, need to have the tool downloaded and installed on a personal computer before making use of its services. It is expected that the proposed method will be compatible with the browsers that are commonly used by users all over the world. The syntactical structure of the links used to lead users to such sites will be considered a starting point. For instance, when a user keys in a group of search terms with the aim of finding web pages that contain information related to the same terms, the tool will come into operation and run through the sites that have been retrieved by the search engine before they are delivered to the user. In doing so, the extension will identify sites whose links contain words that may have a misleading effect on the reader, including those that are characterized by a lot of hyperbole and slang phrases. Such web pages will be flagged as being potential sources of fake news, and the user will be notified before electing to click on either one of them. A visualization of the links and their syntactical structure will help the user understand the decision [5]. Additionally, the tool will also use the number of words associated with the wording used in the titles of the sites for purposes of determining which of them contains false information. A threshold of say eight words will be used as a baseline for categorizing a web page as having correct information, with those whose links containing more than the threshold number of words being classified as potential sources of fake news. The rationale behind this approach is premised on the idea that from a general perspective, clickbaits tend to have considerably longer words than non- clickbaits [14]. It is, therefore, expected that the tool would use the wording as a metric to decide whether a headline can be considered as a potential clickbait. Aside from the syntactic characteristics of the headlines associated with apparent clickbaits, the tool will also monitor how punctuation marks have been used in web pages. In particular, the model will flag sites whose headlines contain extensive usage of exclamation marks and question marks. The links to such web pages will be categorized as potential clickbaits. For instance, a credible site would have a title such as Donald Trump Wins the US Presidential Race! On the other hand, a clickbait would be structured in a manner such as Guess what???? Donald Trump is the Next US President!!!!!!!!!. In such a case, the tool would categorize the former as being a non-clickbait and the latter as being a potential lead to misleading information. In addition, the proposed approach will examine factors associated with individual sites including the bounce rates as a way of determining the veracity (or lack thereof) of the information provided therein. One key characteristic of clickbaits is that they tend to lead readers to web pages containing information that is very different or hardly related to the information highlighted by the link. The result is that a majority of the users end up disappointed, leaving the sites as soon as they have visited it, and resulting in high bounce rates for such web pages [10]. The proposed tool will assess whether a site has a high bounce rate and designate it as a potential source of fake news. Once the algorithm executes, the search engine will release the entire list of results to the user. However, those links whose sites have been noted as being potential sources of misleading information will be highlighted in a manner that

Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 219 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000 5

allows the reader to take notice. Thereupon, the user will be provided with an option of blocking such web pages and having them excluded from the search results in future [3]. It is expected that after using the proposed method for a while, the user will have eliminated a considerable number of clickbaits from the search results retrieved by his or her preferred search engine.

4. METHODOLOGY

The first step was to locate a credible clickbaits database, then compute the attributes and produce the data files for WEKA. That was not easy, therefore, we crawled the web to collect URLs for the clickbaits. We focused on social media web sites that are likely to have more fake news or clickbaits ads or articles, such as: Facebook, Forex and Reddit. The second step, after gathering URLs in a file, a python script computed the attributes from the title and the content of the web pages. Finally, we extracted the features from the web pages. The features are: keywords in Arabic and English, titles that starts with numbers, all caps words, contains question and exclamation marks, if user left the page immediately, and content related to title.

4.1. SCRIPT (PSEUDO CODE)

We had to use WEKA machine learning in order to validate the solution [25]. As WEKA requires specially for- mated input, we used the script below to extract the parameters needed to funiculate WEKA. Ten-fold Cross-validation was used in all experiments.

Algorithm 1 Compute fake news websites attributes

1: Open URL file 2: for each title 3: title starts with number? 1→ output f ile 4: title contains ? and/or ! marks? 1→ output f ile 5: all words are capital in title? 1→ output f ile 6: users left the website after visiting? 1→ output f ile 7: contents have no words from title? 1→ output f ile 8: title contains keywords? NoKeywords→ output f ile 9: end for

4.2. ATTRIBUTES SELECTION

After reading the websites attributes file into WEKA, we rank the attributes based on several algorithms, to choose the most relevant to increase the accuracy and decrease the training time.

• InfoGainAtributeEval evaluates the worth of an attribute by measuring the information gain with respect to the class. In f oGain(Class, Attribute) = H(Class) − H(Class|Attribute). Basically, what it does is measuring how each feature contributes in decreasing the overall entropy. The Entropy, H(X), is defined as follows. H(X) = −sum(Pi ∗ log2(Pi)) with Pi being the probability of the class i in the dataset, and log2 the base 2 logarithm (in WEKA natural logarithm of base e is used, but we take log2). Entropy basically measures the degree of ”impurity”. The closest to 0 it is, the less impurity there is in your dataset. Hence, a good attribute is an attribute that contains the most information, i.e, reduces the most the entropy.

220 Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 6 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000

• CorrelationAttributeEval evaluates the worth of an attribute by measuring the correlation (Pearson’s) between it and the class. Nominal attributes are considered on a value by value basis by treating each value as an indicator. An overall correlation for a nominal attribute is arrived at via a weighted average. So, an indicator for the value of a nominal attribute is a numeric binary attribute that take on the value of 1 when the value occurs in an instance and 0 otherwise.

Table 1 reports the attributes selection results, based on Info Gain and Correlation Attribute, for the tops attributes we use in our tests.

Table 1: Attributes Selection

Attribute Correlation Attribute Eval

Info Gain Attribute Eval

Start with number 0.0768 0.00433 Content have title words 0.775 0.00434 Contain question and excla- mation mark

0.0862 0.00545

All words capital 0.1195 0.104 User left the webpage imme- diately

0.3672 0.12883

Keywords 0.4455 0.27042

4.3. WEKA CLASSIFIERS

The classifier can described as the algorithm that evaluates the given data and provides the end result. WEKA ships with numerous classifiers, we experiments and choose the best performing ones for our dataset.

• BayesNet: Bayes network learning using various search algorithms and quality measures. Bayes Network clas- sifier provides data structures such as network structure, conditional probability distributions, etc., and facilities common to Bayes network learning algorithms such as K2 and B. • Logistic: Class for building and using a multinomial logistic regression model with a ridge estimator. • Random Tree: Class for constructing a tree that considers K randomly chosen attributes at each node. It performs

no pruning and has an option to allow estimation of class probabilities (or target mean in the regression case) based on a hold-out set (backfitting). • NaiveBayaes: Class for a Naive Bayes classifier using estimator classes. Numeric estimator precision values

are chosen based on analysis of the training data. For this reason, the classifier is not an UpdateableClassifier (which in typical usage are initialized with zero training instances).

5. RESULTS

This section presents the performance metrics and discusses the classification results.

5.1. METRICS

Precision is the true positives divided by the predicted positives (the true positives plus the false positives). Mean- while the recall is the rate of the true positives and called also the sensitivity, which is the true positives divided by the true positives plus the false negatives. As for the f-measure, it is the combination of precision and recall, we multiply the precision and recall then divide them to the precision plus the recall and then multiply by two.

Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 221 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000 7

5.2. CLASSIFIERS RESULT

The classifiers are compared based on: Precision, Recall, F-Measure and ROC. Logistic classifier has the highest precision, 99.4% and therefore the best classification quality as shown by Table 2. Logistic and RandomTree classifiers had the best recall that is best sensitivity of 99.3%. The f-measure combines precision and recall, the Logistic and RandomTree classifiers outperformed others at 99.3%. Finally, BayesNet and Naivebayes had the best area under the ROC curve.

Table 2: Classification Results

Classifier Precision Recall F-Measure ROC Bayes Net 94.4% 97.3% 97.2% 100% Logistic 99.4% 99.3% 99.3% 99.5%

RandomTree 99.3% 99.3% 99.3% 97.3% Naive Bayes 98.7% 98.7% 98.6% 100%

6. CONCLUSIONS

Fake news and Clickbaits interfere with the ability of a user to discern useful information from the Internet ser- vices especially when news becomes critical for decision making. Considering the changing landscape of the modern business world, the issue of fake news has become more than just a marketing problem as it warrants serious efforts from security researchers. It is imperative that any attempts to manipulate or troll the Internet through fake news or Clickbaits are countered with absolute effectiveness. We proposed a simple but effective approach to allow users in- stall a simple tool into their personal browser and use it to detect and filter out potential Clickbaits. The preliminary experimental results conducted to assess the method’s ability to attain its intended objective, showed outstanding per- formance in identify possible sources of fake news. Since we started this work, few fake news databases have been made available and we’re currently expanding our approach using R to test its effectiveness against the new datasets.

Acknowledgements

This work was supported by Zayed University Research Office, Research Cluster Award # R17079.

References

References

[1] Abu-Nimeh, S., Chen, T., Alzubi, O., 2011. Malicious and spam posts in online social networks. Computer 44, 23–28. doi:10.1109/MC. 2011.222.

[2] Al Messabi, K., Aldwairi, M., Al Yousif, A., Thoban, A., Belqasmi, F., 2018. Malware detection using dns records and domain name features”, in: International Conference on Future Networks and Distributed Systems (ICFNDS), ACM. URL: https://doi.org/10.1145/3231053. 3231082.

[3] Aldwairi, M., Abu-Dalo, A.M., Jarrah, M., 2017a. Pattern matching of signature-based ids using myers algorithm under mapreduce frame- work. EURASIP J. Information Security 2017, 9. URL: http://dblp.uni-trier.de/db/journals/ejisec/ejisec2017.html# AldwairiAJ17.

[4] Aldwairi, M., Al-Salman, R., 2011. Malurls: Malicious urls classification system, in: Annual International Conference on Information Theory and Applications, GSTF Digital Library (GSTF-DL), Singapore. doi:10.5176/978-981-08-8113-9_ITA2011-29. the best paper award.

[5] Aldwairi, M., Alsaadi, H.H., 2017. Flukes: Autonomous log forensics, intelligence and visualization tool, in: Proceedings of the International Conference on Future Networks and Distributed Systems, ACM, New York, NY, USA. pp. 33:1–33:6. URL: http://doi.acm.org/10. 1145/3102304.3102337, doi:10.1145/3102304.3102337.

222 Monther Aldwairi et al. / Procedia Computer Science 141 (2018) 215–222 8 M. Aldwairi et al. / Procedia Computer Science 00 (2018) 000–000

[6] Aldwairi, M., Hasan, M., Balbahaith, Z., 2017b. Detection of drive-by download attacks using machine learning approach. Int. J. Inf. Sec. Priv. 11, 16–28. URL: https://doi.org/10.4018/IJISP.2017100102, doi:10.4018/IJISP.2017100102.

[7] Balmas, M., 2014. When fake news becomes real: Combined exposure to multiple news sources and political attitudes of inefficacy, alienation, and cynicism. Communication Research 41, 430–454. doi:10.1177/0093650212453600.

[8] Baym, G., Jones, J.P., 2012. News parody in global perspective: Politics, power, and resistance. Popular Communication 10, 2–13. URL: https://doi.org/10.1080/15405702.2012.638566, doi:10.1080/15405702.2012.638566.

[9] Brewer, P.R., Young, D.G., Morreale, M., 2013. The impact of real news about fake news”: Intertextual processes and political satire. In- ternational Journal of Public Opinion Research 25, 323–343. URL: http://dx.doi.org/10.1093/ijpor/edt015, doi:10.1093/ijpor/ edt015.

[10] Chakraborty, A., Paranjape, B., Kakarla, S., Ganguly, N., 2016. Stop clickbait: Detecting and preventing clickbaits in online news media, in: 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), pp. 9–16. doi:10.1109/ ASONAM.2016.7752207.

[11] Chen, Y., Conroy, N.J., Rubin, V.L., 2015. News in an online world: The need for an ”automatic crap detector”, in: Proceedings of the 78th ASIS&T Annual Meeting: Information Science with Impact: Research in and for the Community, American Society for Information Science, Silver Springs, MD, USA. pp. 81:1–81:4. URL: http://dl.acm.org/citation.cfm?id=2857070.2857151.

[12] Conroy, N.J., Rubin, V.L., Chen, Y., 2015. Automatic deception detection: Methods for finding fake news, in: Proceedings of the 78th ASIS&T Annual Meeting: Information Science with Impact: Research in and for the Community, American Society for Information Science, Silver Springs, MD, USA. pp. 82:1–82:4. URL: http://dl.acm.org/citation.cfm?id=2857070.2857152.

[13] Hassid, J., 2011. Four models of the fourth estate: A typology of contemporary chinese journalists. The China Quarterly 208, 813832. doi:10.1017/S0305741011001019.

[14] Lewis, S., 2011. Journalists, social media, and the use of humor on twitter. The Electronic Journal of Communication / La Revue Electronic de Communication 21, 1–2.

[15] Marchi, R., 2012. With facebook, blogs, and fake news, teens reject journalistic objectivity. Journal of Communication Inquiry 36, 246–262. URL: https://doi.org/10.1177/0196859912458700, doi:10.1177/0196859912458700.

[16] Masri, R., Aldwairi, M., 2017. Automated malicious advertisement detection using virustotal, urlvoid, and trendmicro, in: 2017 8th Interna- tional Conference on Information and Communication Systems (ICICS), pp. 336–341. doi:10.1109/IACS.2017.7921994.

[17] Nah, F.F.H., 2015. Fake-website detection tools : Identifying elements that promote individuals use and enhance their performance 1 . introduction.

[18] Pogue, D., 2017. How to stamp out fake news. Scientific American 316, 24–24. doi:10.1038/scientificamerican0217-24. [19] Qbeitah, M.A., Aldwairi, M., 2018. Dynamic malware analysis of phishing emails, in: 2018 9th International Conference on Information and

Communication Systems (ICICS), pp. 18–24. doi:10.1109/IACS.2018.8355435. [20] Riedel, B., Augenstein, I., Spithourakis, G.P., Riedel, S., 2017. A simple but tough-to-beat baseline for the fake news challenge stance detection

task. CoRR abs/1707.03264. URL: http://arxiv.org/abs/1707.03264, arXiv:1707.03264. [21] Rubin, V.L., Chen, Y., Conroy, N.J., 2015. Deception detection for news: Three types of fakes, in: Proceedings of the 78th ASIS&T Annual

Meeting: Information Science with Impact: Research in and for the Community, American Society for Information Science, Silver Springs, MD, USA. pp. 83:1–83:4. URL: http://dl.acm.org/citation.cfm?id=2857070.2857153.

[22] Shu, K., Sliva, A., Wang, S., Tang, J., Liu, H., 2017. Fake news detection on social media: A data mining perspective. SIGKDD Explor. Newsl. 19, 22–36. URL: http://doi.acm.org/10.1145/3137597.3137600, doi:10.1145/3137597.3137600.

[23] Smith, J., Leavitt, A., Jackson, G., 2018. Designing new ways to give context to news stories. https://medium.com/facebook-design/ designing-new-ways-to-give-context-to-news-stories-f6c13604f450.

[24] Spicer, R.N., 2018. Lies, Damn Lies, Alternative Facts, Fake News, Propaganda, Pinocchios, Pants on Fire, Disinformation, Misin- formation, Post-Truth, Data, and Statistics. Springer International Publishing, Cham. pp. 1–31. URL: https://doi.org/10.1007/ 978-3-319-69820-5_1, doi:10.1007/978-3-319-69820-5_1.

[25] of Waikato, U., 2017. Waikato environment for knowledge analysis. URL: https://www.cs.waikato.ac.nz/ml/weka/. [26] Wang, W.Y., 2017. ”liar, liar pants on fire”: A new benchmark dataset for fake news detection. CoRR abs/1705.00648. URL: http:

//arxiv.org/abs/1705.00648, arXiv:1705.00648. [27] Westerman, D., Spence, P.R., Van Der Heide, B., 2014. Social media as information source: Recency of updates and credibility of information.

J. Comp.-Med. Commun. 19, 171–183. URL: http://dx.doi.org/10.1111/jcc4.12041, doi:10.1111/jcc4.12041.

SOCIAL MEDIA

Spotting Misinformation On

Social Media Is Increasingly

Challenging

Peter Suciu Contributor Follow

Aug 2, 2021, 03:56pm EDT

Listen to article 8 minutes

Whether it is about the presidential election, climate change, or Covid-19 vaccines and the

delta … [+] NURPHOTO VIA GETTY IMAGES

Whether it is about the presidential election, climate change, or

Covid-19 vaccines and the delta variant, misinformation continues

to spread rampantly across social media. According to a Pew

New!

Follow this author to improve your

content experience.

Got it!

Research Service study from January, more than eight-in-ten U.S.

adults (86 percent) said they get their news from a smartphone. It

is easy to see why misinformation continues to spread.

While we may expect, even demand that the social platforms crack

down on misinformation, there is little likelihood that Facebook,

Twitter or YouTube will ever stamp out it. One reason is that it

would take full-time policing of virtually all content, but then there

is also the fact that these platforms depend on continued use.

Simply put, misinformation gets clicks.

We see so much misinformation because the platforms have no real

interest in deterring it," explained technology and

telecommunications analyst Roger Entner of Recon Analytics.

"It is really easy and free to join the platform, there is no profit in

deleting the misinformation and preventing provocateurs to post

it," Entner warned.

"Actually, the platforms profit from it because the more outrageous

the content the more people interact with it – this type of

'engagement' is what the platforms are looking for; people reacting

to things. It doesn't matter if it's true or false as long as they

engage," Entner added. "There is also no downside in the

anonymous multimedia world, but even when high profile people

spread lies, there are no repercussions. Everything gets sacrificed

on the altar of monetization through engagement."

Spotting Not Stopping The Misinformation

Since it can't – and likely won't– be stopped, then the best course of

action is spotting it. This may not be as easy as it sounds, because

misinformation is often presented as news and/or fact. In some

cases, it can be wrong by misunderstanding, whilst in other cases it

is misleading by design

MORE FROM FORBES ADVISOR

Best Travel Insurance Companies

By Amy Danise Editor

Best Covid-19 Travel Insurance Plans

By Amy Danise Editor

"Bad information comes in two flavors, unintentional and

intentional. The latter, intentional disinformation, is far more

dangerous," said William V. Pelfrey, Jr., Ph.D., professor in the

Wilder School of Government and Public Affairs at Virginia

Commonwealth University.

"There are many persons who purposefully distribute inaccurate

information in an attempt to influence outcomes, such as an

election," said Pelfrey. "Undermining social confidence during a

time of Covid can disrupt the economy, influence employment, and

negatively impact public health. Some countries, including Russia

and China, have sophisticated disinformation organizations, which

work hard to undermine the United States, thereby elevating other

countries."

ho to believe becomes a problem, too, when there is rampant

contradictory information. This has been made worse as the nation

is so deeply divided, and trust of the "other side" is at an all-time

low.

"Human nature is to simply turn away when things get confusing,"

added Pelfrey.

If the government is telling you to get vaccinated, and social media

is telling you that vaccines and Covid are a hoax, the easiest thing to

do is ignore all of it. That leads to tremendous public health risk,"

he warned. "The smart thing to do is assess the source. What data

support the assertion? Is this fact – based on published, peer-

reviewed research – or opinion? Opinion is often hard to

differentiate from fact as many portray opinion as fact; especially

those soliciting attention, or votes, or your money."

Tips For Spotting Disinformation

There are now ways to quickly spot disinformation, including how

credible the source is – not just on this topic, but past events and

stories as well. In other words, if the source on social media was

wrong in the past that doesn't make them an expert this time

around.

"There are three things that I rely on as my 'go-tos' as quick checks

for patients," said Dr. Donna Gregory, senior lecturer within the

School of Nursing at Regis College. "Who is posting it? What

information are they sharing? What is their intent? In terms of who

is posting it, you want to look at the qualifications and potential for

bias. This is true for both individuals and organizations. For the

information to be credible, the individual or organization should

have credentials that are related to the field. For example, an

infectious disease provider or a healthcare organization that

specializes in infections disease would be credible sources."

The second is what information are they sharing, Gregory added.

Is the information data based on recent studies or information

from science? Or is it a post about one specific case or anecdotal

story," she pondered. "Finally, what is the intent of the person or

group posting? Is the intent to share information based on current

research and science? Are they trying to sell you something? Does

the post generate fear or distrust? Disinformation is more likely to

want to push you in one direction or another, instead of just

sharing information. While it isn't fool proof, considering these

questions together can help you determine if the information is

meant to spread disinformation."

Fact Checking

A common misconception today is that fact checking is often

opinion-based, yet as the very term suggests, it is about checking

the facts. However, fact checking doesn't mean disagreeing with all

the data that counters your argument and only supporting the ones

that agree.

n social media, it is all too common for someone to use a single

source to make or back up an argument while disregarding all the

other facts. That in turn leads to such a spread of misinformation.

If the information comes from a single person, it is opinion—not

fact," said Pelfrey. "My wife saw a friend's post which sounded

scientific, and was written by someone claiming to be an expert,

who turned out to be a part-time pharmacy employee. Not a

scientist."

The problem is made worse as it is makes rounds via social media.

In the John Hughes' film Ferris Bueller's Day Off there is the joke,

"He's sick. My best friend's sister's boyfriend's brother's girlfriend

heard from this guy who knows this kid who's going with a girl who

saw Ferris pass-out at 31 Flavors last night. I guess it's pretty

serious."

That sort of "source" almost flies on social media!

The trappings of expertise are easily fabricated—be skeptical when

reading anything and consider the author's motivation," added

Pelfrey. "Is the author relaying the findings of scientists or are they

trying to push you towards an ideological position?"

And as noted, don't trust one source, especially those that may

seem controversial.

Ferris Bueller – He's SickFerris Bueller – He's Sick

"The main way to check accuracy is to review multiple sources for

the same information," said Gregory. "If you find information about

the delta variant on a news site, can you find this same information

on the CDC website? What about other health care organizations

such as the National Institute of Allergy and Infectious Disease? If

your friend shares information about the vaccines, can you find the

same information on the FDA website or other organizations such

as the World Health Organization? If the information is consistent

across multiple, credible sources than it is likely reliable

information based on evidence."

Follow me on Twitter.

Peter Suciu

I am a Michigan-based writer who has contributed to more than four dozen

magazines, newspapers and websites. I covered the… Read More

Follow

ADVERTISEMENT

Editorial Standards Reprints & Permissions

US Senate

This article is more than 5 years old

Russian deception influenced election due to Trump's support, senators hear Former FBI special agent discusses Russia’s longstanding ‘active measures’ � including the spread of fake news � before Senate intelligence committee

Spencer Ackerman in New York Thu 30 Mar 2017 19.43 BST

Donald Trump’s willingness to embrace Russian disinformation was one of the reasons Russia’s interference in the 2016 election worked, the Senate panel investigating the president’s alleged ties to the country heard on Thursday.

Decades of Russian covert attempts to undermine confidence in western institutions, including planting or promoting false news stories or spreading doubt about the integrity of elections, will accelerate in the future unless the US confronts so-called “active measures”, several experts testified to the Senate intelligence committee.

“Part of the reason active measures have worked in this US election is because the commander-in-chief has used Russian active measures at time [sic] against his opponents,” said Clint Watts of George Washington University’s Center for Cyber and Homeland Security.

Those active measures have migrated online with alacrity in recent years. Watts, a former FBI special agent and army officer who came under personal siege from Russian-backed hackers, told the panel’s first public hearing that social media accounts associated with spreading pro-Russian fake news were visible as far back as 2009.

The expert added that Russia possessed unreleased hacked information on thousands of Americans it could “weaponize” to discredit inconvenient sources. Those and other measures provided Russia with an inexpensive tool to check its wealthier adversaries in the US and Nato, several scholars and former US officials assessed.

When asked why the Russian president, Vladimir Putin, felt the 2016 US election provided the Kremlin an opportunity to intervene – the consensus position of US intelligence agencies – Watts pointed to Trump.

Wittingly or not, Trump and his former campaign chairman, Paul Manafort, embraced and promoted narratives, including false ones, convenient to Russian interests, including a fake story about a terrorist attack on the Turkish airbase at Incirlik used by US forces and baselessly doubting the US citizenships of Barack Obama and Ted Cruz.

“On 11 October, President Trump stood on stage and cited what appears to be a fake news story from [the Russian propaganda outlet] Sputnik News that disappeared from the internet. He denies the intel from the United States about Russia. He claimed that the election could be rigged – that was the number one theme pushed by RT, Sputnik News,” Watts testified.

The first public hearing by the Senate intelligence committee into its Trump-Russia investigation featured none of the partisan rancor that has defined its counterpart in the House of Representatives.

The Republican chairman, Richard Burr of North Carolina, himself a Trump ally during the campaign, pledged a “thorough, independent and non-partisan review” of evidence potentially tying Trump to Russia, a commitment echoed by the top panel Democrat, Mark Warner of Virginia.

“If we politicize this process, our efforts will likely fail,” Burr said at the outset of the hearing.

I write from Ukraine, where I've spent much of the past six months, reporting on the build-up to the conflict and the grim reality of war. It has been the most intense time of my 30-year career. In December I visited the trenches outside Donetsk with the Ukrainian army; in January I went to Mariupol and drove along the coast to Crimea; on 24 February I was with other colleagues in the Ukrainian capital as the first Russian bombs fell.

This is the biggest war in Europe since 1945. It is, for Ukrainians, an existential struggle against a new but familiar Russian imperialism. Our team of reporters and

The House of Representatives’ inquiry has taken a back seat to the public drama surrounding its leader, the California Republican Devin Nunes. Nunes is under Democratic pressure to recuse himself after his promotion of information provided to him at the White House and cancellations of public hearings prompted allegations of covering up for Trump. The New York Times on Thursday reported that two White House staffers aided Nunes’s effort last week, Ezra Cohen-Watnick of the National Security Council and Michael Ellis, an attorney who formerly worked on Nunes’s committee.

Nunes’s spokesman, Jack Langer, said: “As he’s stated many times, Chairman Nunes will not confirm or deny speculation about his source’s identity, and he will not respond to speculation from anonymous sources.”

The Senate’s hearings have not yet featured key testimony from serving intelligence and law enforcement officials, so they have not yet encountered the same high-stakes pressure as the House’s, which last week saw the leadership of the FBI and NSA publicly reject Trump’s evidence-free allegation that Obama placed him under surveillance.

But in what Burr and Warner promoted on Wednesday as a sign of their seriousness, Burr assigned seven committee staffers to the inquiry. That would be more than Burr’s predecessor, the California Democrat Dianne Feinstein, had on the vast review of CIA torture, though that inquiry quickly lost GOP support.

Urging a response to Russian interference in the election, Watts said the US approach to Russia was provocatively ambiguous.

“I’m not sure what our policy or stance is with regards to Russia at this point in the United States. I think that’s the number one thing we’ve got to figure out, because that will shape how they interface with us,” he told senators.

More on this story

Eric Trump said family golf courses attracted Russian funding, author claims

8 May 2017

Trump�Russia investigation reignites as Senate asks aides to hand over notes

5 May 2017

Mike Flynn under formal investigation by Pentagon over payments from Russia

27 Apr 2017

editors intend to cover this war for as long as it lasts, however expensive that may prove to be. We are committed to telling the human stories of those caught up in war, as well as the international dimension. But we can't do this without the support of Guardian readers. It is your passion, engagement and financial contributions which underpin our independent journalism and make it possible for us to report from places like Ukraine.

If you are able to help with a monthly or single contribution it will boost our resources and enhance our ability to report the truth about what is happening in this terrible conflict.

Thank you.

Continue Remind me in August

Luke Harding Foreign correspondent

Single Monthly Annual

CA$5 per month CA$10 per month Other

More from Headlines

Roe v Wade � Anger as protests continue across US after supreme court overturns law

3h ago

Dom Phillips � Murdered British journalist laid to rest in Brazil

3h ago

Hajj � British Muslim travel agencies in uproar over Saudi changes

4h ago

Credit: Cristina Spanò

Information Overload Helps Fake News Spread, and Social Media Knows It Understanding how algorithm manipulators exploit our cognitive vulnerabilities empowers us to fight back

By:Filippo Menczer, Thomas Hills

December 1, 2020

Consider Andy, who is worried about contracting COVID-19. Unable to read all the articles he sees on it, he relies on trusted friends for tips. When one opines on Facebook that pandemic fears are overblown, Andy dismisses the idea at first. But then the hotel where he works closes its doors, and with his job at risk, Andy starts wondering how serious the threat from the new virus really is. No one he knows has died, after all. A colleague posts an article about the COVID “scare” having been created by Big Pharma in collusion with corrupt politicians, which jibes with Andy's distrust of government. His Web search quickly takes him to articles claiming that COVID-19 is no worse than the flu. Andy joins an online group of people who

h b f b i l id ff d fi d hi lf ki lik f h “Wh

have been or fear being laid off and soon finds himself asking, like many of them, “What pandemic?” When he learns that several of his new friends are planning to attend a rally demanding an end to lockdowns, he decides to join them. Almost no one at the massive protest, including him, wears a mask. When his sister asks about the rally, Andy shares the conviction that has now become part of his identity: COVID is a hoax.

This example illustrates a minefield of cognitive biases. We prefer information from people we trust, our in-group. We pay attention to and are more likely to share information about risks—for Andy, the risk of losing his job. We search for and remember things that fit well with what we already know and understand. These biases are products of our evolutionary past, and for tens of thousands of years, they served us well. People who behaved in accordance with them—for example, by staying away from the overgrown pond bank where someone said there was a viper—were more likely to survive than those who did not.

Modern technologies are amplifying these biases in harmful ways, however. Search engines direct Andy to sites that inflame his suspicions, and social media connects him with like- minded people, feeding his fears. Making matters worse, bots—automated social media accounts that impersonate humans—enable misguided or malevolent actors to take advantage of his vulnerabilities.

Compounding the problem is the proliferation of online information. Viewing and producing blogs, videos, tweets and other units of information called memes has become so cheap and easy that the information marketplace is inundated. Unable to process all this material, we let our cognitive biases decide what we should pay attention to. These mental shortcuts influence which information we search for, comprehend, remember and repeat to a harmful extent.

The need to understand these cognitive vulnerabilities and how algorithms use or manipulate them has become urgent. At the University of Warwick in England and at Indiana University Bloomington's Observatory on Social Media (OSoMe, pronounced “awesome”), our teams are using cognitive experiments, simulations, data mining and artificial intelligence to comprehend the cognitive vulnerabilities of social media users. Insights from psychological studies on the evolution of information conducted at Warwick inform the computer models developed at Indiana, and vice versa. We are also developing analytical and machine-learning aids to fight social media manipulation. Some of these tools are already being used by journalists, civil-society organizations and individuals to detect inauthentic actors, map the spread of false narratives and foster news literacy.

Th l f i f i h d i i i f l ' i A N b l

I N F O R M A T I O N O V E R L O A D

The glut of information has generated intense competition for people's attention. As Nobel Prize–winning economist and psychologist Herbert A. Simon noted, “What information consumes is rather obvious: it consumes the attention of its recipients.” One of the first consequences of the so-called attention economy is the loss of high-quality information. The OSoMe team demonstrated this result with a set of simple simulations. It represented users of social media such as Andy, called agents, as nodes in a network of online acquaintances. At each time step in the simulation, an agent may either create a meme or reshare one that he or she sees in a news feed. To mimic limited attention, agents are allowed to view only a certain number of items near the top of their news feeds.

Running this simulation over many time steps, Lilian Weng of OSoMe found that as agents' attention became increasingly limited, the propagation of memes came to reflect the power- law distribution of actual social media: the probability that a meme would be shared a given number of times was roughly an inverse power of that number. For example, the likelihood of a meme being shared three times was approximately nine times less than that of its being shared once.

Credit: “Limited individual attention and online virality of low-quality information,” By Xiaoyan Qiu et al., in Nature Human Behaviour, Vol. 1, June 2017

This winner-take-all popularity pattern of memes, in which most are barely noticed while a few spread widely, could not be explained by some of them being more catchy or somehow more valuable: the memes in this simulated world had no intrinsic quality. Virality resulted purely from the statistical consequences of information proliferation in a social network of agents with limited attention. Even when agents preferentially shared memes of higher quality, researcher Xiaoyan Qiu, then at OSoMe, observed little improvement in the overall quality of those shared the most. Our models revealed that even when we want to see and share high-quality information, our inability to view everything in our news feeds inevitably leads us to share things that are partly or completely untrue.

C i i bi l h bl I f db ki di i

Cognitive biases greatly worsen the problem. In a set of groundbreaking studies in 1932, psychologist Frederic Bartlett told volunteers a Native American legend about a young man who hears war cries and, pursuing them, enters a dreamlike battle that eventually leads to his real death. Bartlett asked the volunteers, who were non-Native, to recall the rather confusing story at increasing intervals, from minutes to years later. He found that as time passed, the rememberers tended to distort the tale's culturally unfamiliar parts such that they were either lost to memory or transformed into more familiar things. We now know that our minds do this all the time: they adjust our understanding of new information so that it fits in with what we already know. One consequence of this so-called confirmation bias is that people often seek out, recall and understand information that best confirms what they already believe.

This tendency is extremely difficult to correct. Experiments consistently show that even when people encounter balanced information containing views from differing perspectives, they tend to find supporting evidence for what they already believe. And when people with divergent beliefs about emotionally charged issues such as climate change are shown the same information on these topics, they become even more committed to their original positions.

Making matters worse, search engines and social media platforms provide personalized recommendations based on the vast amounts of data they have about users' past preferences. They prioritize information in our feeds that we are most likely to agree with—no matter how fringe—and shield us from information that might change our minds. This makes us easy targets for polarization. Nir Grinberg and his co-workers at Northeastern University recently showed that conservatives in the U.S. are more receptive to misinformation. But our own analysis of consumption of low-quality information on Twitter shows that the vulnerability applies to both sides of the political spectrum, and no one can fully avoid it. Even our ability to detect online manipulation is affected by our political bias, though not symmetrically: Republican users are more likely to mistake bots promoting conservative ideas for humans, whereas Democrats are more likely to mistake conservative human users for bots.

Credit: Filippo Menczer

In New York City in August 2019, people began running away from what sounded like gunshots. Others followed, some shouting, “Shooter!” Only later did they learn that the blasts came from a backfiring motorcycle. In such a situation, it may pay to run first and ask questions later. In the absence of clear signals, our brains use information about the crowd to infer appropriate actions, similar to the behavior of schooling fish and flocking birds.

Such social conformity is pervasive. In a fascinating 2006 study involving 14,000 Web-based volunteers, Matthew Salganik, then at Columbia University, and his colleagues found that when people can see what music others are downloading, they end up downloading similar

M h l i l d i “ i l” i hi h h ld h

S O C I A L H E R D I N G

songs. Moreover, when people were isolated into “social” groups, in which they could see the preferences of others in their circle but had no information about outsiders, the choices of individual groups rapidly diverged. But the preferences of “nonsocial” groups, where no one knew about others' choices, stayed relatively stable. In other words, social groups create a pressure toward conformity so powerful that it can overcome individual preferences, and by amplifying random early differences, it can cause segregated groups to diverge to extremes.

Social media follows a similar dynamic. We confuse popularity with quality and end up copying the behavior we observe. Experiments on Twitter by Bjarke Mønsted and his colleagues at the Technical University of Denmark and the University of Southern California indicate that information is transmitted via “complex contagion”: when we are repeatedly exposed to an idea, typically from many sources, we are more likely to adopt and reshare it. This social bias is further amplified by what psychologists call the “mere exposure” effect: when people are repeatedly exposed to the same stimuli, such as certain faces, they grow to like those stimuli more than those they have encountered less often.

Credit: Jen Christiansen; Source: Dimitar Nikolov and Filippo Menczer (data)

Such biases translate into an irresistible urge to pay attention to information that is going viral—if everybody else is talking about it, it must be important. In addition to showing us items that conform with our views, social media platforms such as Facebook, Twitter, YouTube and Instagram place popular content at the top of our screens and show us how many people have liked and shared something. Few of us realize that these cues do not provide independent assessments of quality.

In fact, programmers who design the algorithms for ranking memes on social media assume that the “wisdom of crowds” will quickly identify high-quality items; they use popularity as a proxy for quality. Our analysis of vast amounts of anonymous data about clicks shows that all

l f i l di h i d i f i ll i f i

platforms—social media, search engines and news sites—preferentially serve up information from a narrow subset of popular sources.

To understand why, we modeled how they combine signals for quality and popularity in their rankings. In this model, agents with limited attention—those who see only a given number of items at the top of their news feeds—are also more likely to click on memes ranked higher by the platform. Each item has intrinsic quality, as well as a level of popularity determined by how many times it has been clicked on. Another variable tracks the extent to which the ranking relies on popularity rather than quality. Simulations of this model reveal that such algorithmic bias typically suppresses the quality of memes even in the absence of human bias. Even when we want to share the best information, the algorithms end up misleading us.

Most of us do not believe we follow the herd. But our confirmation bias leads us to follow others who are like us, a dynamic that is sometimes referred to as homophily—a tendency for like-minded people to connect with one another. Social media amplifies homophily by allowing users to alter their social network structures through following, unfriending, and so on. The result is that people become segregated into large, dense and increasingly misinformed communities commonly described as echo chambers.

At OSoMe, we explored the emergence of online echo chambers through another simulation, EchoDemo. In this model, each agent has a political opinion represented by a number ranging from −1 (say, liberal) to +1 (conservative). These inclinations are reflected in agents' posts. Agents are also influenced by the opinions they see in their news feeds, and they can unfollow users with dissimilar opinions. Starting with random initial networks and opinions, we found that the combination of social influence and unfollowing greatly accelerates the formation of polarized and segregated communities.

Indeed, the political echo chambers on Twitter are so extreme that individual users' political leanings can be predicted with high accuracy: you have the same opinions as the majority of your connections. This chambered structure efficiently spreads information within a community while insulating that community from other groups. In 2014 our research group was targeted by a disinformation campaign claiming that we were part of a politically motivated effort to suppress free speech. This false charge spread virally mostly in the conservative echo chamber, whereas debunking articles by fact-checkers were found mainly in the liberal community. Sadly, such segregation of fake news items from their fact-check reports is the norm.

S i l di l i i i I l b d R b J i ll

E C H O C H A M B E R S

Social media can also increase our negativity. In a recent laboratory study, Robert Jagiello, also at Warwick, found that socially shared information not only bolsters our biases but also becomes more resilient to correction. He investigated how information is passed from person to person in a so-called social diffusion chain. In the experiment, the first person in the chain read a set of articles about either nuclear power or food additives. The articles were designed to be balanced, containing as much positive information (for example, about less carbon pollution or longer-lasting food) as negative information (such as risk of meltdown or possible harm to health).

The first person in the social diffusion chain told the next person about the articles, the second told the third, and so on. We observed an overall increase in the amount of negative information as it passed along the chain—known as the social amplification of risk. Moreover, work by Danielle J. Navarro and her colleagues at the University of New South Wales in Australia found that information in social diffusion chains is most susceptible to distortion by individuals with the most extreme biases.

Even worse, social diffusion also makes negative information more “sticky.” When Jagiello subsequently exposed people in the social diffusion chains to the original, balanced information—that is, the news that the first person in the chain had seen—the balanced information did little to reduce individuals' negative attitudes. The information that had passed through people not only had become more negative but also was more resistant to updating.

A 2015 study by OSoMe researchers Emilio Ferrara and Zeyao Yang analyzed empirical data about such “emotional contagion” on Twitter and found that people overexposed to negative content tend to then share negative posts, whereas those overexposed to positive content tend to share more positive posts. Because negative content spreads faster than positive content, it is easy to manipulate emotions by creating narratives that trigger negative responses such as fear and anxiety. Ferrara, now at the University of Southern California, and his colleagues at the Bruno Kessler Foundation in Italy have shown that during Spain's 2017 referendum on Catalan independence, social bots were leveraged to retweet violent and inflammatory narratives, increasing their exposure and exacerbating social conflict.

Information quality is further impaired by social bots, which can exploit all our cognitive loopholes. Bots are easy to create. Social media platforms provide so-called application programming interfaces that make it fairly trivial for a single actor to set up and control thousands of bots. But amplifying a message, even with just a few early upvotes by bots on

i l di l f h R ddi h h i h b l i

R I S E O F T H E B O T S

social media platforms such as Reddit, can have a huge impact on the subsequent popularity of a post.

At OSoMe, we have developed machine-learning algorithms to detect social bots. One of these, Botometer, is a public tool that extracts 1,200 features from a given Twitter account to characterize its profile, friends, social network structure, temporal activity patterns, language and other features. The program compares these characteristics with those of tens of thousands of previously identified bots to give the Twitter account a score for its likely use of automation.

In 2017 we estimated that up to 15 percent of active Twitter accounts were bots—and that they had played a key role in the spread of misinformation during the 2016 U.S. election period. Within seconds of a fake news article being posted—such as one claiming the Clinton campaign was involved in occult rituals—it would be tweeted by many bots, and humans, beguiled by the apparent popularity of the content, would retweet it.

Bots also influence us by pretending to represent people from our in-group. A bot only has to follow, like and retweet someone in an online community to quickly infiltrate it. OSoMe researcher Xiaodan Lou developed another model in which some of the agents are bots that infiltrate a social network and share deceptively engaging low-quality content—think of clickbait. One parameter in the model describes the probability that an authentic agent will follow bots—which, for the purposes of this model, we define as agents that generate memes of zero quality and retweet only one another. Our simulations show that these bots can effectively suppress the entire ecosystem's information quality by infiltrating only a small fraction of the network. Bots can also accelerate the formation of echo chambers by suggesting other inauthentic accounts to be followed, a technique known as creating “follow trains.”

Some manipulators play both sides of a divide through separate fake news sites and bots, driving political polarization or monetization by ads. At OSoMe, we recently uncovered a network of inauthentic accounts on Twitter that were all coordinated by the same entity. Some pretended to be pro-Trump supporters of the Make America Great Again campaign, whereas others posed as Trump “resisters”; all asked for political donations. Such operations amplify content that preys on confirmation biases and accelerate the formation of polarized echo chambers.

Understanding our cognitive biases and how algorithms and bots exploit them allows us to better guard against manipulation. OSoMe has produced a number of tools to help people

d d h i l bili i ll h k f i l di l f

C U R B I N G O N L I N E M A N I P U L A T I O N

understand their own vulnerabilities, as well as the weaknesses of social media platforms. One is a mobile app called Fakey that helps users learn how to spot misinformation. The game simulates a social media news feed, showing actual articles from low- and high- credibility sources. Users must decide what they can or should not share and what to fact- check. Analysis of data from Fakey confirms the prevalence of online social herding: users are more likely to share low-credibility articles when they believe that many other people have shared them.

Another program available to the public, called Hoaxy, shows how any extant meme spreads through Twitter. In this visualization, nodes represent actual Twitter accounts, and links depict how retweets, quotes, mentions and replies propagate the meme from account to account. Each node has a color representing its score from Botometer, which allows users to see the scale at which bots amplify misinformation. These tools have been used by investigative journalists to uncover the roots of misinformation campaigns, such as one pushing the “pizzagate” conspiracy in the U.S. They also helped to detect bot-driven voter- suppression efforts during the 2018 U.S. midterm election. Manipulation is getting harder to spot, however, as machine-learning algorithms become better at emulating human behavior.

Apart from spreading fake news, misinformation campaigns can also divert attention from other, more serious problems. To combat such manipulation, we have recently developed a software tool called BotSlayer. It extracts hashtags, links, accounts and other features that co- occur in tweets about topics a user wishes to study. For each entity, BotSlayer tracks the tweets, the accounts posting them and their bot scores to flag entities that are trending and probably being amplified by bots or coordinated accounts. The goal is to enable reporters, civil-society organizations and political candidates to spot and track inauthentic influence campaigns in real time.

These programmatic tools are important aids, but institutional changes are also necessary to curb the proliferation of fake news. Education can help, although it is unlikely to encompass all the topics on which people are misled. Some governments and social media platforms are also trying to clamp down on online manipulation and fake news. But who decides what is fake or manipulative and what is not? Information can come with warning labels such as the ones Facebook and Twitter have started providing, but can the people who apply those labels be trusted? The risk that such measures could deliberately or inadvertently suppress free speech, which is vital for robust democracies, is real. The dominance of social media platforms with global reach and close ties with governments further complicates the possibilities.

One of the best ideas may be to make it more difficult to create and share low-quality information. This could involve adding friction by forcing people to pay to share or receive information. Payment could be in the form of time, mental work such as puzzles, or

i i f f b i i A d i h ld b d lik

Scientific American is part of Springer Nature, which owns or has commercial relations with thousands of scientific publications (many of them can be found at www.springernature.com/us). Scientific American maintains a strict policy of editorial independence in reporting

developments in science to our readers.

© 2022 SCIENTIFIC AMERICAN, A DIVISION OF SPRINGER NATURE AMERICA, INC.

ALL RIGHTS RESERVED.

microscopic fees for subscriptions or usage. Automated posting should be treated like advertising. Some platforms are already using friction in the form of CAPTCHAs and phone confirmation to access accounts. Twitter has placed limits on automated posting. These efforts could be expanded to gradually shift online sharing incentives toward information that is valuable to consumers.

Free communication is not free. By decreasing the cost of information, we have decreased its value and invited its adulteration. To restore the health of our information ecosystem, we must understand the vulnerabilities of our overwhelmed minds and how the economics of information can be leveraged to protect us from being misled.

This article was originally published with the title "The Attention Economy" in Scientific American 323, 6, 54-61

(December 2020)

doi:10.1038/scientificamerican1220-54

View This Issue