Freedom of choice and network "echo effects" of information consumption

Leontyev Gleb Dmitrievich PhD in Philosophy Associate Professor, Department of General Philosophy, Kazan (Volga region) Federal University 35 Kremlevskaya str., Kazan, 420008, Russia, Republic of Tatarstan leontyeval@icloud.com Other publications by this author

Leontieva Ludmila Stanislavovna PhD in Philosophy Associate Professor, Department of State and Municipal Administration, Kazan (Volga Region) Federal University 420008, Russia, Republic of Tatarstan, Kazan, Kremlevskaya str., 18 lsl3@yandex.ru Other publications by this author

The dynamics of digitalization and mediatization of society generates both technological and socio-communicative effects. Online social media platforms are technologies for virtualizing social reality. "Humanity," according to J. Baudrillard decided to clone his physicality and his possessions in another universe, different from the previous one," in which "simple informativeness, calculability, and computability reign" ^[1]. Virtuality simulatively replaces reality, intensely absorbing the life world of everyday life. The speed, scale, and depth of information penetration through social network interactions is comparable to the virulence of an infectious agent. Following the recent pandemic, in line with epidemiological analogies, infection with viral information is associated with the immunological resistance of the body, with the ability of the pathogen to overcome protective barriers. That is, the "virulence" of virtual objects depends both on the properties, "packaging" of information products, and on the susceptibility of an individual and a community to mass information.

 The ability of virtuality to create the illusion of its own reality can be traced by the example of the client-oriented formation of a "unique information universe for each of us" ^[2]. This figurative formulation belongs to the American political Internet activist Eli Pariser and means, the term he introduced, "bubble filter" - a bubble of algorithms, filters, an information bubble. Externally, the information bubble looks like "an ecosystem of familiar applications, selected platforms and news aggregators, as well as a set of "subscriptions" and personal "representations" (pages) on social networks" ^{[3, p. 13]}. The listed information intermediaries by selective information supplies save the human psyche from information overload, but the same selection algorithms narrow the field of view, providing information that strengthens the preferences of the individual and confirms his rooted ideas. As a result, the effect of an acoustic wave reflected from the walls of the bubble chamber is created, which the recipient accepts with satisfaction on the principle of "quietly with himself" and with his own kind. Professor Cass Sunstein warned about this effect more than two decades ago in his book "Echo Chambers" ^[4]. The intensive development of social network interactions, information redundancy in conjunction with information consumerism actualize the problem of side effects of online communications. Therefore, the purpose of this article is to identify factors affecting the effectiveness of content filtering algorithms on social platforms, on the user's ability to make an informed choice in the process of information consumption.

Echo techniques in the psychology of communication appeared long before the era of social networks, they are based on the effect of enhancing exposure through verbal and non-verbal repetition. In online communications, the echo effect is created by selective strategies of network behavior, which are based on the human need for self-identification, acceptance and confirmation of one's own feelings and reflections from the outside, in increasing one's strength and resources through belonging to a community. Therefore, the echo chamber is a "complex phenomenon of spontaneous formation of stable forms of network communication" ^[5], i.e. a situationally conditioned space of like-mindedness. According to the Global Digital 2023 report, 64.4% of the world's population has Internet access, 60% use social networks, and the average time spent on networks is 4 minutes out of every 10 minutes online. 30% of respondents named social networks as a source of receiving news on the Internet, search engines – 25%, aggregators – 8%. The influence of intermediary platforms is growing, and media sites and applications are used less and less: in 2018 - 32%; in 2023 – 22% ^[6]. Compared to direct access to familiar news sites, the algorithms of the platforms increase the degree of diversity of the sources of information received. At the same time, search engines are configured based on the preferences of users who make their choice based on the suggestions of algorithms. "You find yourself trapped in an "I" loop" discovered by Eli Pariser more than a decade ago: "Your identity shapes your media consumption, but then the media shapes your beliefs and areas of interest" ^{[2, p. 139]}. The mechanisms of media personalization are combined with basic cognitive distortions, such as "selective receptivity" and "confirmation bias", which migrated from the offline environment to the online space together with the media subjects, since they allow users to avoid differences that lead to psychological discomfort.

A quantitative assessment of the importance of the motive of "confirmation of beliefs" in the formation of demand for news was given by American researchers (F. Chopra, I. Haaland, Chris. Roth) ^[7] based on several large-scale experiments with voters. The willingness of people to subscribe to the newsletter was studied depending on the presentation of news content: from a politically impartial approach with reliable facts to a biased, politically biased (right, left). Key preference parameters showed that respondents reduce their demand for "biased news" only if this type of bias is incompatible with their own political beliefs. Noting the ability of biased news to increase political polarization and the growth of populism, the researchers noted the presence of a consumer compromise: the equivalence of the motive of information reliability and the motive of "confirmation bias" ^{[7, p. 29]}.  Consequently, the objectivity, evidence, and accuracy of the social facts provided to the audience does not guarantee media success. The basis of modern information and economic relations, the so-called "supervisory" (Shoshana Zuboff) ^[8], "platform" (Nick Srnicek) ^[9], "communicative" (Jody Dean) ^[10] capitalism is a marketing adjustment to a specific consumer, optimization of the technological chain of creating a journalistic product as a sought–after product. The delivery of individually relevant information, that is, the possibility of personalization or targeting is feasible, thanks to the digital footprint of the user. For example, analyzing likes/dislikes, taking into account previous user actions with similar content, posts, and search history allow you to create a digital portrait that includes not only a list of preferences, but also predictive scenarios of media consumer behavior.

The peculiarities of consumer behavior caused by the considered psychological mechanisms of "bias" and "selectivity" are not only the consumption of thematically defined, value-semantic content, but also the orientation of social interactions on an online platform. Taking into account this specificity, scientists from the University of Oslo (M. Cinelli, G. de F. Morales, A. Galeazzi, M. Starnin) studied the communications of more than one million active users on four social media platforms, in total - more than one hundred million unique fragments on controversial, socially significant issues. The presence/absence of echo chambers was assessed according to two criteria: homophilia in interactions on a specific topic and bias in the dissemination of information from like-minded sources. The results of a study published in 2021 showed that the degree of segregation in news consumption varies depending on the platform, but in general, "aggregation of users into homophile clusters dominates online interactions" ^[11], i.e. social media platforms and news feed algorithms contribute to the emergence of echo chambers.

The filter bubble and the echo chamber are, at first glance, identical concepts. However, the way they are formed has differences, which Dr. Richard Fletcher points out: "echo chambers can be the result of filtration or other processes, whereas a filter bubble is the result of algorithmic filtration" ^[12]. If in the first case users consciously filter information, then in the second this task is performed by the algorithms of the platforms, i.e. the criterion for the difference of the concepts under consideration is a method of selection. The approach proposed by the Oxford researcher of digital news consumption models can be defined as socio-technological. In the Russian scientific literature, along with the socio-technological one, a cognitive-communicative, logical-semantic perspective of analysis is proposed. Professor V. A. Bazhanov notes that "the difference between the two communicative spaces lies in a certain tolerance for dissent in the case of echo bubbles and an intolerant attitude towards it in the case of echo chambers" ^{[13, p. 155]}.  If "epistemic echo bubbles" are formed due to the similarity of people's views (in a broad sense "on life", including political) and the emotions accompanying these views" ^{[13, p. 152]}, then the "echo chamber" is a closed communicative formation, "strengthening, expanding a kind of epistemic control over the state of minds and forming special structures for countering and exposing the authoritative opinions of the opposite side" ^{[13, p. 156]}. When these approaches are schematically superimposed, revealing the differences between "echo chambers and bubbles", it turns out that the machine algorithm for the formation of "echo bubbles" is more tolerant than consciously constructed filters of "echo chambers" with a one-sided distortion of reality perception, with dogmatics and intra-group cognitive attitudes. Despite the conditionality of this mental operation, its result emphasizes the importance of the thoughts and actions of the user himself, contrary to the idea of the omnipotence of filtering algorithms. If the user's installations and selective communications are able not only to form data filtering algorithms, but also to enhance their work, then we can assume the user's ability to neutralize or bypass filters.

Any machine filtering algorithms or psychological traps are manipulations that suppress, but do not cancel, freedom of choice. A meaningful choice involves the activity of the mind on the basis of private interest, expediency and logical ordering, taking into account real conditions, the relationship of goals, means and the external environment. The ability of modern users to rationalize behavior in the network media environment, reducing the effectiveness of the impact of algorithmic filtering are due to the following factors:

Firstly, the predictive abilities of artificial intelligence can fail due to difficulties in interpreting user behavior: inconsistency in views and preferences, argumentative tricks, logical errors, aesopian language, "social automatism" ^[14] of network interactions. To recognize the "pragmatics of human information manipulation" ^[15], a machine algorithm could benefit from competencies that include the Popper principle of falsification - refutation of one's own ideas.  The differences in the status and role self-presentation of users are due to the fact that they are not limited to one social platform, but create accounts in different social networks, on websites and online services, or use unregistered access to materials, to discussions on open forums. "The content promoted or hidden by the algorithms of the platforms is mostly created by people and therefore is not without contradictions" ^[16]. For example, the clickbait phenomenon worsens the user experience, but the author of the content gets an advanced position in the output due to the inadequate reaction of the system, i.e. the algorithms do not calculate the lack of semantic connection between the title and the content of the material. In addition to interpreting the "subjectivisms" of users, algorithms need "additional education" to understand the social context and, according to Professor Y. Harari, to correct mistakes made by programmers at the level of "subconscious biases" ^[17].

Secondly, the adoption of the social and value attitudes of the network community does not entail the obligatory uniformity of thoughts and actions of all its members in all spheres of network existence. For example, a study of one hundred and fifteen virtual communities of the Russian social network Vkontakte allowed experimenters to conclude that not only echo chambers function, but also rational confrontational communication. "The blocking of critical accounts did not exceed 30% of the number of communities included in the sample for each of the ideological groups" ^{[18, p. 82]}.

Thirdly, the conceptualization of negative network effects, according to media theorist Axel Bruns, may have a conjunctural motivation on the part of some media politicians and technopessimists ^{[16, pp. 16-18]}: either in order to tighten the regulation of social media, or in order to reduce the independent activity of users who believe in strict dependence on algorithmic services of online platforms.

Fourthly, social technologies of influencing mass consciousness have existed since ancient times, as well as means of conscious counteraction. The preservation of "human agency" to the extent necessary for the user is achievable by the efforts of the user himself. This goal is served not only by information technology, but also by cognitive and psychological methods of protection.

In terms of information technology, self-preservation measures available to the average user in the algorithmic media space involve disabling recommendations in video hosting settings, disabling application access to the microphone and abandoning the voice assistant (audio privacy); abandoning automatic sorting of publications on networks, from the personalization function in the search service, as well as systematic cleaning of the history of search queries and views video content.

By focusing on the "degree of agency" demanded by the user himself, we emphasize not only the negative, but also the positive sides of personalized content filtering. The duality is explained by the convenience of receiving recommendations selected by algorithms. A machine hint or, in R. Avenarius's terminology, work on the "principle of least waste of effort" is optimal in a standard situation, for example, when searching for information of a reference, descriptive nature. The independent choice of consumed content (for example, subscription) and following the recommendations of algorithms in network practice, in our opinion, do not exclude each other, but complement each other, are applied according to the situation.  The default settings are a convenient option in the "architecture of choice" ^[19], says Professor Cass Sunstein, giving an example of using a navigator, whose suggestions can be rejected if desired. In his book "The Illusion of Choice", the professor, defending the right to active choice, at the same time considers the "choice not to choose" as one of the ways to manifest free will. In the case of conformal behavior, decision-making is delegated to experts, politicians and algorithms. Such "default settings" are akin to the psychological mechanism of stereotyping, when the mode of "saving thinking" is turned on and you can act according to a template, based on indirect experience and ready-made, simplified ideas.

The user's need for an in-depth analysis of a specific problem requires rational behavior in the information sphere as well as in any other area of life. Information technology methods of protection in combination with cognitive and psychological methods help a person to preserve his own subjectivity. The formed echo chamber can be significantly reduced or abolished only by "the disappearance of the event itself that led to its appearance" ^{[13, p.159]}. The risk of being in the "echo chamber" is minimized at the initial "launch" of a critical attitude to information, since the level of confirmation bias is reduced. This is evidenced by the results of an experiment by American researchers ^[20] who studied the impact of fake news on social networks. The participants of the experiment, bachelors from the Indiana University School of Business, had to express their attitude to news materials. The trust was aroused by articles that coincided with their own point of view ("confirmation bias"), but when creating the so-called "user rating" of articles, the participants took a critical position. The task was to publicly assess the reliability of information about which the participants of the experiment had no personal experience, which led to doubt and caution in information exchange. Consequently, the researchers concluded, the news does not become viral when the critical filter is activated, since it reduces the demand for biased content.

 Critical attitude to information is associated, on the one hand, with the degree of accessibility and diversity of sources, and, on the other hand, with the level of openness of the user himself to messages that do not fit into the idealized picture of individual perception. The breadth of views, the willingness to understand the opponent's position contribute to the conscious acceptance of alternative arguments through doubt, comparative analysis, the search for similarities and comprehension of the essence of differences. At the same time, the argumentative meaningfulness of the conclusion does not exclude a state of psychological discomfort: from a feeling of one's own cognitive insufficiency to a crisis of self-identification. The consistent implementation of the installation to go beyond the echo-resonance bubble activates the internal locus of user control, changing the perspective of self-assembly and the content of information needs.

Incentives for reasonable information consumption arise on the basis of: 1) lack of available information for the implementation of certain actions; 2) contradictions between available information and information coming from outside; 3) contradictions between media news, theoretical, ideological innovations and everyday practice. As a result of the awareness of inconsistencies, a "search image" of the information need is formed, i.e. an idea of one's own information gaps in the amount of information that is in demand to solve the problem. An independently identified trend of information search leads to the desired goal, bypassing externally set, algorithmized "landmarks". With the hypothetical transparency of the filters of online platforms, the user has the opportunity to interact with the algorithms under his control, setting a conscious trajectory of his own development and search. In both cases, the lack of information determination on the part of algorithms takes the user outside the information-closed community.

Thus, the collection, storage, dissemination and monetization of information about possible consumer behavior stimulate the formation of filter bubbles. The trend towards separation between online communities based on consumed content is confirmed by research experiments on various social platforms. The ability of artificial intelligence to self-study increases the attractiveness and accuracy of personalized recommendations. However, it is premature to talk about the totality of self-contained echo chambers, about the impenetrability of the boundaries of network communities. Communication and reasonable information consumption, awareness of assessments and conclusions increase the degree of adequacy of perception of reality in all its contradictory diversity, reduce the likelihood of rooted stay in a comfortable information bubble.

Delegating decision-making to artificial intelligence, i.e. "choosing not to choose", or preserving and strengthening one's own subjectivity, contrary to smart algorithms and neural network filters, is a conscious, situational choice of the user himself, his self-organization in the process of responsible information consumption.