This chapter aims at presenting the main technological instruments and initiatives in the field of combating online disinformation, playing a significant role in the general architecture of the handbook, helping the target group understand the phenomenon of disinformation in a comprehensive manner, by acknowledging which are the weapons one can use in order to avoid becoming a victim of false content. Therefore, the chapter will try to display the main features of different technical solutions to combat online disinformation, also aims at developing the readers’ digital skills by exploring.

Therefore, this chapter will focus on (1) identifying and briefly presenting the main directions that can be followed in order to combat the negative effects of disinformation through the use of technology; (2) defining the main concepts used within this chapter, so as to establish a common theoretical background for the approached directions; (3) briefly presenting the main technological solutions available on the market for both raising the level of awareness of the public opinion (at large) towards the effects of online disinformation and digitally training the public opinion to avoid becoming a victim of the online disinformation phenomenon; and (4) examining the limitations that the technological solutions have when it comes to detecting disinformation. Therefore, the chapter will define concepts such as media literacy, serious games etc., presenting their applicability in the field of disinformation. This section will also briefly analyse core technological tools used for detecting and countering disinformation and discuss their potential and limitations, providing a brief analysis on the human rights impact of such tools.

The result of this chapter can be structured into three main areas: (1) a first part dedicated to establishing a common understanding of the key concepts that will be developed and employed in the whole chapter; (2) a compiled list of one of the most known applications and other technological solutions to counteract and combat (to a certain level) the effects of online disinformation; (3) the main limitations and challenges posed by technological solutions to countering online disinformation.  

Digital competences addressed

1.1 Browsing, searching and filtering data, information and digital content;
1.2 Evaluating data, information and digital content;
1.3 Managing data, information and digital content;
2.1 Interacting through digital technologies;
2.2 Sharing through digital technologies;
2.3 Engaging citizenship through digital technologies;
2.6 Managing digital identity. 

5.1 Combating the effects of disinformation in the online environment

Alexandra Anghel. Ana Ćuća. Aitana Radu


Recently, both technological developments and advances registered in the sector of Internet of Things and social networks have created the premises for the expansion of the noxious effects of the disinformation phenomena. Even though disinformation is not considered a recent phenomenon, it gained widespread attention from governments worldwide as a result of its applicability in the online environment, throughout the use of social media. Therefore, topics such as fake news, disinformation, propaganda have experienced a resurgence of interest in nowadays societies, resulting from the generalized concerns around the widespread negative effects and impact of disinformation on public opinion and public events (Sharma, et al, 2019, 2).
Furthermore, the rise of ubiquitous misinformation, disinformation, deepfakes, and post-truth raised also increasing concerns on the role of the Internet and social media in current democratic societies. Given its fast and widespread diffusion, disinformation imposes not only an individual and societal cost, but can also determine economic losses or national security risks (Fraga-Lamas & Fernandez-Carames, 2020, 53). In addition, the features of online communication, such as the speed and scope at which false information can be disseminated in the online environment, led to an increased potential of the people to deceive throughout the usage of computer-mediated communication channels, aspect that can produce major changes on financial markets, as well as political scenes (Fuller, Biros, & Wilson, 2009).
Taking into account the above-mentioned aspects, it is true to say that technology has created the means for the expansion of the disinformation phenomenon, social media becoming one of the main sources of information for the population at large, as well as an important source of false content and digital deception. However, technology can also play an essential role in combating the effects of online disinformation and propaganda and in containing the expansion processes of these now defined security issues. This chapter aims, therefore, at presenting the main technological instruments and initiatives in the field of combating online disinformation, in addition to the previous sections that present the techniques employed by different disinformation processes. This chapter plays a significant role in the general architecture of the handbook, helping the target group understand the phenomenon of disinformation in a comprehensive manner, by acknowledging which are the weapons one can use in order to avoid becoming a victim of false content, as well as developing their digital skills by exploring different technical solutions to combat online disinformation.
This section is structured into two main parts: (1) a part dedicated to establishing a common understanding of the key concepts that will be developed and employed in the whole chapter, and (2) a compiled list of one of the most known applications and other technological solutions to counteract and combat (to a certain level) the effects of online disinformation.

Main research questions addressed

●  What is the role of technological tools in fighting disinformation?
●  Which are the main technological initiatives in in terms of combating disinformation and what is their potential and limitations?

Combating the effects of disinformation in the online environment – setting the scene

Since 2016, the world is actively witnessing the rise of disinformation across different media platforms. The reason behind it lies in the intersection of several factors. Firstly, since 2016, there is increasing online propaganda disseminated through hyper partisan news sites that use disinformation as a business model for generating profit. During the 2016 US elections, the number of “news sites'' that fabricated pro-Trump news skyrocketed. Moreover, these news sites were not restricted to the US, but could even be found in remote parts of the world. For example, over 100 pro-Trump websites were registered in the small town of Veles in North Macedonia, which produced viral fake stories promoting Trump (Posetti, 2018). In 2017, Facebook confirmed that Russia spent over $100,000 to finance ads which spread polarizing views on different societal topics, such as immigration, race and LGBTQI rights, all of which were topics of discussion during the 2016 presidential elections (Menn & Ingram, 2017). Secondly, politicians are increasingly using propaganda terms to frame political issues, instead of employing a fact-based approach. In 2017 alone, former US president Donald Trump made over 1,999 false or misleading claims (Kessler et al., 2019). Every time Trump, or any other politician repeats misleading claims, even when these have been proven to be untrue by the media, they still have a significant impact on public trust. Thirdly, the technological advancement in the field of advertising algorithms and social media platforms enabled the creation of partisan camps and polarized crowds. Search engine optimization, personalized social media feeds, and micro-targeted advertising allow polarized crowds to consume content that confirms their prior beliefs (Kessler et al., 2019).

Moreover, there is no significant difference between exposure to fake news in Europe vs. the US. According to the results of a 2018 public consultation organized by the European Commission, out of 2986 participants, 97% of them claimed to have been exposed to disinformation, 38% on a daily basis, and 32% on a weekly basis. The majority of participants (74%) believed that social media facilitates the spread of disinformation (European Commission, 2018).

The issue of disinformation has become even more prominent in the context of the Ukrainian war. Both the previous Donbas conflict in Ukraine and the current war show how countries such as Russia are extensively engaging in information warfare via social media platforms.

One potential answer to this growing trend of online propaganda and disinformation is the use of fact-checking by the media, think tanks and/or individual experts (see also section 3.4 for more information on fact-checking). However, fact-checking can be a lengthy process that includes several stages, such as:

  1. Identification: Includes constant media monitoring and constant monitoring of political sources. Given the amount of news that are published daily, this stage also includes prioritization of claims who urgently need to go through the fact-checking procedure;
  2. Verification: Includes checking the identified claim against an already existing fact-check as well as checking the information against official sources. In some occasions, the verification stage will also include credibility sourcing.
  3. Correction: Includes flagging the false information, providing additional data to contextualize provided information, and publishing the corrected news (Alphilippe et al., 2019).

Given the lengthiness of the fact-checking procedure and the amount of disinformation that is continuously produced and shared, there is a growing need for using automated fact-checking tools and other tech-driven solutions. Development of such tools allows interested groups (fact checkers, media, academia, policymakers) to understand the most important news/claims that need to be fact-checked, to timely react if someone is sharing disinformation and, most importantly, to detect disinformation that is starting to circulate.

Apart from fact-checking, there is a growing need for detecting social media bots which manipulate online discussions and are often used for spreading disinformation and manipulating narratives (Alphilippe et al., 2019). Whereas technology can be used to amplify disinformation on social networks either through the creation and promotion of disinformation or through the use of social media bots, tech-driven solutions are also leading the way in the fight against disinformation.

The growth of computer-mediated communication though the usage of social media registered during the last decade, as well as the changes produced in the terminology specific of the disinformation phenomena led to an evolution of the core nature and characteristics of the problem itself (Sharma, et al, 2019, p. 5). In order to better understand the way in which technology can be employed to counteract the negative effects of disinformation, it is important to acknowledge the factors that allow fake news and disinformation to spread at both individual and social level.

In consequence, in addition to the factors addressed in chapters 2, when referring to the individual level, the literature in the field has shown that the inability of a person to discern in an accurate manner false content from real one led to an uninterrupted process of sharing and believing of false information disseminated on social media (Sharma, et al, 2019, 5). As an example, a survey conducted by the international research, data and analytics group YouGov in 2017 on 1684 British adults who were required to analyze the credibility of six individual news stories (half of which were false and the other half true) found that only 4% of the individuals had the capacity to correctly identify them (Channel 4, 2017). The inability to distinguish false content from the real one was attributed to ideological biases and cognitive abilities. In addition, authors Gordon Pennycook and David G. Rand presented in their study the positive correlation between propensity for analytical thinking and the ability to differentiate false from true content (Pennycook & Rand, 2019). However, authors Hunt Allcott and Matthew Gentzkow showed that there are differences in the way in which people perceive information available on social media, generated by the time allotted for consuming media content, their level of education and their age (with higher educated, older people being more accurate in forming perceptions of information) (Allcott & Gentzkow, 2017, 228).

Adjacent to cognitive abilities, ideological priors can also play an important role in the process of information consumption. Naive realism (which refers to the individual tendency to trust more easily in information that is aligned with his/her own views), confirmation bias (individuals tend to select and prefer to receive only that information which confirms their existing views, rejecting any piece of information that contravene their points of view), and normative influence theory (individuals have the tendency to disseminate and consume socially safe options as a preference for social acceptance and affirmation) are generally considered important elements in the perception and dissemination of fake news and disinformation (Shu, Sliva, Wang, Tang, & Liu, 2017, 24).

As the social level, the core of social media and collaborative information sharing on online platforms provides a supplementary dimension to disinformation and fake news, generally known as the echo chamber effect (Shu, Sliva, Wang, Tang, & Liu, 2017, 225). The principles of naive realism, confirmation bias, and normative influence theory stated above, describe the need of individuals to search, consume, and disseminate information that is in alignment with their own viewpoints and ideologies, developing, in consequence, the tendency to establish and develop connections with ideologically similar individuals (social homophily). Therefore, social media algorithms focus on customizing recommendations (algorithmic personalization) by suggesting content that better fits an individual’s preferences, as well as by recommending connections to persons that share similar beliefs (Sharma, et al, 2019, 5).

Both social homophily and algorithmic personalization contribute to the development of echo chambers and filter bubbles, wherein individuals get less exposure to conflicting viewpoints and become isolated in their own information sphere (Garimella, Gionis, Parotsidis, & Tatti, 2017, 4663). The existence of echo chambers can increase the chances of survival and continuous dissemination of fake news, aspect that can be explained by the phenomena of social credibility and frequency heuristic. The concept of social credibility indicates that people’s perception of credibility of a piece of information tends to increase if others also perceive it as credible (especially in those cases when the credibility of the source of information cannot be tested), and the frequency heuristic concept defines the tendency to grant a higher level of credibility to a piece of information to which an individual is exposed multiple times (Shu, Sliva, Wang, Tang, & Liu, 2017, 24).

Given these factors, as well as those referred to in chapters 1 and 2, one can conclude that technology has equipped the general public with new and more developed capabilities to consume different media contents, from streaming video to reading niche blogs and news sites. In addition, technology also developed for people worldwide the habit of trusting the transparency of the content they consume each day from social media, without questioning or evaluating the source of information (Fowler, 2022). However, technology did not only create the premises for the expansion of online disinformation, but it also allowed the development of solutions to combat the negative effects of the above mentioned phenomena.

The majority of tech-driven solutions are relying on machine learning. The attractiveness of machine learning in the context of targeting and combating disinformation arises from the fact that machine learning models can recognize novel cases and react to them, based on prior learning. The possibility of continuous improvement of machine learning models, makes them seem like an effective tool to address the always-evolving world of disinformation. In this context, the next section of the chapter will map the main assessment methods used for developing technological instruments to combat online disinformation (identified based on an analysis on the existing literature) and current efforts from the Machine Learning (ML) community to fight against the threats posed by the disinformation phenomenon at both individual and social level.

Assessment methods – a review

In order to set the framework for a better understanding of the main solutions identified in the domain of combating the negative effects of online disinformation, it is essential to define the assessment methods that the most majority of technological solutions are based on.

Therefore, the literature in the domain divides the methods, which emerged from various domains, using disparate techniques, into two main categories (see Conroy, Rubin, & Chen, 2015):

(1) Linguistic approaches – focus on extracting and analyzing the content of deceptive messages in order to associate language patterns with deception.

More specifically, this type of approach is aimed at identifying "leakages” in the content of the message analyzed by measuring the frequency and patterns of pronouns, conjunction, negative emotion word usage and so on (Feng & Hirst, 2013).

The methods associated with this category are, as follows: 

Linguistic approaches 

Focus on extracting and analyzing the content of deceptive messages in order to associate language patterns with deception. More specifically, this type of approach is aimed at identifying „leakages” in the content of the message analyzed by measuring the frequency and patterns of pronouns, conjunction, negative emotion word usage and so on (Feng & Hirst, 2013).

The methods associated with this category are, as follows:

Data representation

One of the simplest methods of representing texts is the “bag of words” approach, which regards each word as a single, equally significant unit. In the bag of words approach, individual words or “n- grams” (multiword) frequencies are aggregated and analyzed to reveal cues of deception. However, by relying on isolated n-grams, often divorced from useful context information, any resolution of ambiguous word sense remains non-existent (Conroy, Rubin, & Chen, 2015, 2). 

Deep syntax

Since the analysis of word use is sometimes not enough in predicting deception, deeper language structures (syntax) have been developed as a complementary solution. Deep syntax analysis is implemented through Probability Context Free Grammars (PCFG), that transforms sentences in a set of rewrite rules (a parse tree) to describe syntax structure (e.g. noun and verb phrases), which are in turn rewritten by their syntactic constituent parts (Feng, Banerjee, & Choi, 2012). The final set of rewrites produces a parse tree with a certain probability assigned. The method is used to distinguish rule categories (lexicalized, un-lexicalized, parent nodes etc.) for deception detection with 85-91% accuracy (depending on the rule category used) (Conroy, Rubin, & Chen, 2015, 2).

However, used alone, this method might not be sufficiently capable of identifying deception, therefore studies often combine this approach with other linguistic or network analysis techniques (Feng, Banerjee, & Choi, 2012) (Feng & Hirst, 2013).

Semantic analysis

This approach extends the n-gram plus syntax model by incorporating profile compatibility features, showing the addition significantly improves classification performance (Feng & Hirst, 2013). This method uses the principle of aligning profiles and the description of the writer’s personal experience, in order to assess veracity based on compatibility scores:

1. compatibility with the existence of some distinct aspect (e.g. an art museum near a mentioned hotel);
2. compatibility with the description of some general aspect, such as location or service. In this case, the prediction of falsehood is shown to be at approximately 91% accurate (Conroy, Rubin, & Chen, 2015, 2

Rhetorical Structure and Discourse Analysis

A method used to achieve the description of discourse, by identifying the instances of rhetoric relations between linguistic elements (Rubin & Lukoianova, 2014)

Network approaches 

Focus on the usage of network properties and behavior to complement content-based approaches that rely on deceptive language and leakage cues to predict deception.

The methods associated with this category are, as follows (Conroy, Rubin, & Chen, 2015, 3)

Linked data 

An approach that leverages an existing body of collective human knowledge in order to assess the truth of new statements. The method is based on querying existing knowledge networks, or publicly available structured data (e.g. the Google Relation Extraction Corpus (GREC)). The structured data network of entities is connected through a predicate relationship. This particular method can help develop the applicability of fact-checking methods (Conroy, Rubin, & Chen, 2015, 3)

Social network behaviour

Besides content analysis, the use of metadata and telltale behavior of questionable sources can be examined. This method focuses on compiling the inclusion of hyperlinks or associated metadata to establish veracity assessments. As an example, centering resonance analysis (CRA), is a model of network-based text analysis, representing the content of large sets of texts by identifying the most important words that link other words in the network (Conroy, Rubin, & Chen, 2015, 3-4)

Machine Learning (ML) solutions to online disinformation

Complementary to the methods presented above, the detection of false information and fake news can be performed by analyzing multiple types of digital content: images, text data, network data, as well as the credibility degree of the author/source and its reputation (Choraś, Demestichas, Giełczyk, & Herrero, 2021, 1-2).

A survey conducted by a team of researchers from the University of Albany on the solutions to address fake news detection through text-analysis and mainstream fake news datasets showed that the state-of-the-art approaches for combating the effects of disinformation through detection can be clustered into five main categories, depending on the methods they use (Parikh & Atrey, 2018, 438)

Linguistic features based methods – which extract key linguistic features from fake news and false information, as follows:

Ngrams: unigrams and bigrams are extracted from the matrix of words in a certain story. These are most often stored as TFIDF (Term Frequency Inverse Document Frequency) values for information retrieval. TFIDF refers to a numerical statistic that is intended to reflect how important a word is to the document that it is used in (Parikh & Atrey, 2018, 438);

② Punctuation: using punctuation in an article can help the algorithms for fake news detection to make the difference between false and truthful texts. In this case, the punctuation feature collects eleven types of punctuation, implemented through this detection (Parikh & Atrey, 2018, 438);

Psycho-linguistic features: which use the LIWC lexicon (Linguistic Inquiry and Word Count) in order to pick out appropriate proportions of words, allowing the system to determine the tone of the language (e.g. positive/negative emotions), statistics of the text (e.g. word counts), part-of-speech category (e.g. articles, nouns, verbs) (Pérez-Rosas, Kleinberg, Lefevre, & Mihalcea, 2018, 5);

④ Readability: includes the extraction of content features such as the number of characters, complex words, long words, number of syllables, word types, and number of paragraphs (Pérez-Rosas, Kleinberg, Lefevre, & Mihalcea, 2018, 5);

Syntax: focuses on extracting a set of features based on CFG (context-free grammar), which are heavily dependent on lexicalized production rules combined with their parent and grandparent nodes. Functions in this set are also encoded in TFIDF for information retrieval purposes (Parikh & Atrey, 2018, 439). 

Deception modelling based models – use the two theoretical techniques described below to convert texts into a set of rhetorical relations connected in a hierarchical tree and identifying the results of rhetorical structure relations:

Rhetorical Structure Theory (RST): focuses on capturing the logic of a story in terms of functional relations created amongst different meaningful text units, describing, at the same time, a hierarchical structure for each story (Mann & Thompson, 1988). In accordance with the findings of the authors Victoria Rubin, Nadia Conroy and Yimin Chen (Rubin, Conroy, & Chen, 2015), empirical research confirmed in the last decades that writers tend to emphasise certain parts of their papers so as to express in a more evident manner the main ideas expressed in that article. In this context, the RST uses rhetorical connections to identify, in a systematic manner, the emphasized parts of a text (Parikh & Atrey, 2018, 439);

② Vector Space Modeling (VSM): used to identify the rhetorical structure relations in the sets resulted after the application of RST. VSM helps at interpreting every news text as vectors in high dimensional space, aspect that requires for the extracted text to be modeled in an appropriate manner for the application of various computational algorithms (Rubin, Conroy, & Chen, 2015). In this context, each dimension of a certain vector space refers to the number of rhetorical relations in a complete set of news reports, representation which provides a simple explanation of a vector space, making it available for further analysis (Rubin & Lukoianova, 2014) (Parikh & Atrey, 2018, 439).     

Clustering based models – a known method to compare and contrast a large amount of data. For example, the gCLUTO package (Graphical CLUstering TOolkit) runs a large number of data set and sorts a small number of clusters using agglomerative clustering with the k-nearest neighbor approach, clustering similar news reports based on the normalized frequency of relations (Rubin, Conroy, & Chen, 2015);

Predictive modelling based methods – develop the ability to make predictions about previously unseen news pieces on the results of a logistic regression process (Rubin, Conroy, & Chen, 2015);

Content cues based models – a model based on the ideology of what journalists like to write for users and what are the preferences of users in terms of reading (choice gap), that leverages two different analyses: (I) lexical and semantic levels of analysis (automated methods can be used to extract stylometric features of the text, such as subjective terms, word length etc., further used to differentiate between journalistic formats) and (II) syntactic and pragmatic levels of analysis (the pragmatic function of headlines invokes reference to forthcoming parts in the discourse by making reference to forthcoming parts in the news story. This analysis also covers measuring news sites which have more share activity compared to sites that substantially produce more news content) (Parikh & Atrey, 2018, 440).

Types of digital content that can be analyzed in order to detect fake news by using automatic instruments

Types of digital content that can be analyzed in order to detect fake news by using automatic instruments (Choraś, Demestichas, Giełczyk, & Herrero, 2021, p. 2)

Considering the solutions briefly described above, we can conclude that there are many Machine Learning initiatives developed in order to combat the negative effects of the online disinformation phenomenon. However, the next section of this chapter will focus on the most common ones, that can be understood by the general public, that does not possess a technological background: (1) serious games – used in order to familiarize the general public with the processes specific to online disinformation and determine it to acknowledge the effects that can be produced by its online behavior and (2) natural language processing tools and (3) social media bot-detecting tools – both used in order to facilitate the process of analyzing the trust level of content disseminated on social media.

However, the chapter does not display a comprehensive analysis of all the technological solutions currently available for combating the effects of online disinformation and propaganda but will only focus on those that were most advertised and that require common, not-specialized skills in order to be used with effective results. Therefore, this chapter will only describe those technological solutions that help to develop digital skills, namely the ones most widely used, examining their potential, but also limitations and societal impact.

Lastly, this chapter offers some recommendations for the further improvement of the use of such tools, especially in what concerns compliance with existing human rights standards. However, the chapter will not include a user manual for each example offered but will focus on presenting the main features of the technological solutions and their benefits (the actual cycle of steps that must be followed in order to exploit the features of each application will be developed in the second deliverable of the project). 

Inspiring practices, projects, interventions in the field

In addition to the above-mentioned examples, in terms of fake news detection initiatives (with emphasis on rumors), both industry and the scientific community have registered efforts to identify and develop solutions, ranging from research projects (ongoing or already implemented) to fully-fledged applications.

The most notable and well-known examples have been collected in the table below (the data has been extracted from Zubiaga, Aker, Bontcheva, Liakata, & Procter, 2018, 24-26).

Name of the initiative


Brief description


A 3-year research project funded by the European Commission, implemented during the period of 2014-2017, focusing on the study of natural language processing techniques for dealing with rumour detection and resolution


A data-driven, real-time, web-based rumour tracker, which tracks automatically social media mentions of URLs’ associated rumours; however, the identification of rumours and selection of URLs associated with those has not been automated and still requires human input. It was part of a research project led by Craig Silverman, partnering with the Tow Center for Digital Journalism at Columbia University, which focuses on how unverified information and rumour are reported in the media (Silverman, 2015)



A 1-year research project that was implemented in 2014, funded by Google. Its main objective was to build a tool to aid journalists in finding posts that spread or correct a particular rumour on Twitter, by trying to identify the size of the audiences that those posts have reached. Further details on the rumour detection system developed in this project were published in (Zhao, Resnick, & Mei, 2015)


A project in the Social Informatics Lab at Wellesley College. Twitter Trails was developed as an interactive, web-based tool that allows users to conduct an investigation on the origin and propagation characteristics of a rumour and its refutation, if applicable, on Twitter. Visualisations of burst activity, propagation timeline, RT, and co-retweeted networks help its users trace the spread of a story. It collects relevant tweets and automatically answers several important questions regarding a rumour: its originator, burst characteristics, propagators, and main actors according to the audience. In addition, this tool computes and reports the rumour’s level of visibility and, as an example of the power of crowdsourcing, the audience’s skepticism toward it, which correlates with the rumour’s credibility.



A framework that designs, adopts and implements multiple visualizations and modelling tools that can be mixed to identify rumour contents and analyze participant activity, either within a rumour, or across different rumours. This approach helps analysts in drawing hypotheses regarding rumour propagation (Dang, Smit, Mod'h, & Minghim, 2016)



A 3-year research project funded by the European Commission, focusing on studying the role of social media to crisis management. In this context, the project studied the adverse use and reliability of social media, including the impact of rumours (Scifo & Baruh, 2013)



A 3-year research project funded by the European Commission that studied the
use of social sensors for security assessments and proactive emergencies management, dealing also with crowdsourced annotation of rumours (McCreadie, Macdonald, & Ounis, 2015) 


A platform for the collection, detection and analysis of online misinformation and its related fact-checking efforts


A 3-year project (2013–2016) funded by the European Commission, that aimed at verifying social media content from a journalistic and enterprise perspective, with a focus especially on image verification. The project results were represented by a number of publications on journalistic verification practices concerning social media (Brandtzaeg, Lüders, Spangenberg, Rath-Wiggins, & Følstad, 2016), social media verification approaches (Andreadou, Papadopoulos, Apostolidis, Krithara, & Kompatsiaris, 2015), and approaches to track down the location of social media users (Middleton & Krivcovs, 2016).


A Horizon 2020 project, funded by the European Commission (2017-2020), with the target to build a platform providing services to detect, authenticate, and check the reliability and accuracy of newsworthy video files and video content spread via social media.


A collaborative verification project implemented by First Draft and Google News Lab, in collaboration with a number of newsrooms in France, with the objective to fight misinformation (mainly focusing on the French presidential election)


An online database by the French news organization Le Monde, that allows user to check the reliability of news sites


A project aiming to perform live fact-checking. The demo application shows check-worthy claims identified by the system for the 2016 U.S. election and it allows the user to input their own text to find factual claims


A real-time, web-based system developed to assess the credibility of content posted on Twitter. The system does not determine the veracity of stories, but it provides a credibility rating (scored 1 to 7) for each tweet in the Twitter timeline.


A global coalition of stakeholders advancing the fight against digital disinformation and malicious deepfakes

Project application 

aASearch Engine which presents insights and analytics on various topics of your choice by combining the power of Crowd-Sourcing with next-generation technologies like Blockchain, Machine Learning and AI.

It is clear that there are multiple technologies which play/will play an important role in detecting and combating disinformation. They protect democracies and their citizens from unlawful interference in their internal processes and shed light on the mechanisms used to manipulate public opinions. This positive impact is not without costs. As explained in this deliverable, many of these technologies are still underdevelopment and thus subject to many limitations. In addition to the high error rate, there are also cases where their use can have a negative impact on human rights. In order to avoid this, stronger emphasis needs to be placed on understanding technological limitations, introducing privacy-by-design and privacy-by-default approaches in their developments as well as carrying out a constant review of ways in which their design can be improved in order to mitigate potential risks. In addition to this, it is important to ensure that the regulatory framework manages to keep the pace with technological developments, by introducing the necessary safeguards.

1. Allcott, H., & Gentzkow, M. (2017). Social Media and Fake News in the 2016 Election. Journal of Economic Perspective, 31(2), 211-236.
2. Andreadou, K., Papadopoulos, S., Apostolidis, L., Krithara, A., & Kompatsiaris, Y. (2015). Media REVEALr: A social multimedia monitoring and intelligence system for web multimedia verification. Pacific-Asia Workshop on Intelligence and Security Informatics.
3. Brandtzaeg, P. B., Lüders, M., Spangenberg, J., Rath-Wiggins, L., & Følstad, A. (2016). Emerging journalistic verification practices concerning social media. Journalism Practice, 10(3), 323–342.
4. Choraś, M., Demestichas, K., Giełczyk, A., & Herrero, Á. (2021). Advanced Machine Learning techniques for fake news (online disinformation) detection: A systematic mapping study. Applied Soft Computing Journal(101), 1-15.
5. Conroy, N. J., Rubin, V. L., & Chen, Y. (2015). Automatic Deception Detection: Methods for Finding Fake News. Proceedings of the Association for Information Science and Technology, 52(1), 1-4.
6. Feng, V. W., & Hirst, G. (2013). Detecting Deceptive Opinions with Profile Compatibility. Nagoya, Japan: Proceedings of the Sixth International Joint Conference on Natural Language Processing.
7. Feng, S., Banerjee, R., & Choi, Y. (2012). Syntactic Stylometry for Deception Detection. Jeju Island, Korea: Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers).
8. Fowler, G. (2022). Fake News, Its Impact and How Tech Can Combat Misinformation. Retrieved January 11, 2023, from https://www.forbes.com/sites/forbesbusinessdevelopmentcouncil/2022/08/22/fake-news-its-impact-and-how-tech-can-combat-misinformation/?sh=308d6bab354f
9. Fraga-Lamas, P., & Fernandez-Carames, T. M. (2020). Fake News, Disinformation, and Deepfakes: Leveraging Distributed Ledger Technologies and Blockchain to Combat Digital Deception and Counterfeit Reality. IT Professional, 22(2), 53-59.
10. Fuller, C., Biros, D. P., & Wilson, R. L. (2009). Decision Support for Determining Veracity Via Linguistic-Based Cues. Decision Support Systems, 46(3), 695-703.
11. Garimella, K., Gionis, A., Parotsidis, N., & Tatti, N. (2017). Balancing Information Exposure in Social Networks. (pp. 4663-4671). Advances in Neural Information Processing Systems.
12. Kessler, G., Rizzo, S., Kelly, M. (2019, December 16). President Trump has made 15,413 false or misleading claims over 1,055 days. The Washington Post. https://www.washingtonpost.com/politics/2019/12/16/president-trump-has-made-false-or-misleading-claims-over-days/
13. Menn, J., Ingram, D. (2017, September 6). Facebook says likely Russian-based operation funded U.S. ads with political message. Reuters. https://www.reuters.com/article/us-facebook-propaganda-idUSKCN1BH2VX
14. Middleton, S. E., & Krivcovs, V. (2016). eoparsing and Geosemantics for Social Media: Spatiotemporal Grounding of Content Propagating Rumors to Support Trust and Veracity Analysis during Breaking News. ACM Transactions on Information Systems, 34(3), 1-26.
15. Parikh, S. B., & Atrey, P. K. (2018). Media-Rich Fake News Detection: A Survey. 2018 IEEE Conference on Multimedia Information Processing and Retrieval.
16. Pennycook, G., & Rand, D. G. (2019). Who falls for fake news? The roles of bullshit receptivity, overclaiming, familiarity, and analytic thinking. Journal of Personality, 1-63.
17. Posetti, J., Matthews, A. (2018). A short guide to the history of ’fake news’ and disinformation. International Centre for Journalists. https://www.icfj.org/sites/default/files/2018-07/A%20Short%20Guide%20to%20History%20of%20Fake%20News%20and%20Disinformation_ICFJ%20Final.pdf
18. Rubin, V. L., & Lukoianova, T. (2014). Truth and Deception at the Rhetorical Structure Level. Journal of the Association for Information Science and Technology, 66(5), 905-917.
19. Sharma, K., Qian, F., Jiang, H., Ruchansky, N., Zhang, M., & Liu, Y. (2019). Combating Fake News: A Survey on Identification and Mitigation Techniques. ACM Transactions on Intelligent Systems and Technology, 10(3), 1-42.
20. Shu, K., Sliva, A., Wang, S., Tang, J., & Liu, H. (2017). Fake News Detection on Social Media: A Data Mining Perspective. ACM SIGKDD Explor. Newslett, 19(1), 22-36.
21. Zubiaga, A., Aker, A., Bontcheva, K., Liakata, M., & Procter, R. (2018). Detection and Resolution of Rumours in Social Media: A Survey. ACM Computing Surveys, 51(2), 1-36. 

Co-funded by European Commission Erasmus+
University of Malta
University Rey Juan Carlos
Logo New Strategy Center

Project: DOMINOES Digital cOMpetences INformatiOn EcoSystem  ID: 2021-1-RO01-KA220-HED-000031158
The European Commission’s support for the production of this publication does not constitute an endorsement of the contents, which reflect the views only of the authors, and the Commission cannot be held responsible for any use which may be made of the information contained therein.

Ivan, Cristina; Chiru, Irena; Buluc, Ruxandra; Radu, Aitana; Anghel, Alexandra; Stoian-Iordache, Valentin; Arcos, Rubén; Arribas, Cristina M.; Ćuća, Ana; Ganatra, Kanchi; Gertrudix, Manuel; Modh, Ketan; Nastasiu, Cătălina. (2023). HANDBOOK on Identifying and Countering Disinformation. DOMINOES Project https://doi.org/10.5281/zenodo.7893952