The Tyranny of Perceived Opinion: Freedom and Information in the Era of Big Data - Reading GVPT 241 for 9/23/25
Introduction
The digital age presents a paradox: unprecedented access to information, yet challenges in effectively utilizing it.
This is due to sophisticated means of curating and delivering information through technologies such as algorithms, big data, and artificial intelligence.
The author examines how this curation leads to phenomena like filter bubbles and echo chambers, where individuals primarily encounter information aligning with pre-existing beliefs.
This tendency, partly stemming from selective exposure, where individuals seek information that supports their existing views and avoid contradictory, unpleasant information, creates a significant problem.
When information, suggestions, and portrayals create expectations, they become leading and potentially detrimental to liberty.
The combination of these technologies and human nature poses a threat to liberty by undermining individuality, autonomy, and the foundations of liberal society.
Drawing an analogy to Alexis de Tocqueville's fear of the "tyranny of the majority," the author suggests we should now fear the "tyranny of algorithms and perceived opinion."
The article focuses on how liberty is threatened by contemporary information handling, starting with psychological aspects like selective exposure.
It argues that despite information abundance, people might be less exposed to undesirable, uncomfortable, or provocative ideas, potentially leading to less liberal societies and hindering the full development of individuality.
Special attention is given to how big data and algorithms exacerbate these issues by delivering information to us and delivering information from us to others.
The author employs theories of expectation, arguing they are crucial additions to debates on the non-neutral, manipulative, and shaping nature of big data and algorithms.
Traditional political philosophy provides a framework for understanding the threat to liberty, especially when liberty is considered a broader concept, encompassing social as well as political phenomena, rather than merely freedom from intentional physical interference.
This broader understanding highlights the threat to individuality, autonomy, and liberal society, focusing on perceived opinions, expectations, and the renewed relevance of a tyranny of opinion.
The work aligns with calls to analyze how technology impacts the good society where liberty is a core value, emphasizing the importance of political philosophy.
While primarily focused on individual effects, the issues also bear significant societal importance, connecting to increased polarization, changing capitalism, and the health of liberal democracy (referencing Sunstein for implications for deliberative democracy).
The author asserts that the "tyranny of opinion" feared by Mill and Tocqueville may be more oppressive now within the digitally curated landscape.
Background and Terminology: Big Data Information
Our Era: Often referred to as the era of Big Data.
Lainie's Definition of Big Data:
Volume: The enormous amount of information available today.
Velocity: The speed at which data is generated and gathered.
Variety: The structural heterogeneity of the data.
Additional Characteristics of Big Data:
Organic: Everything is collected, faithfully reflecting the real world more than traditional data.
Global: A worldwide phenomenon.
Artificial Intelligence (AI): A term describing machines capable of tasks requiring intelligence, from simple industrial robots to complex decisions like loan approvals or bail determinations.
Machine Learning (ML): Systems that improve performance on a given task over time through experience, rather than explicit human training or control.
Synergy: The core subject of this work is the world created when algorithms learn from big data.
The Problem is Not the Technology Itself (Zuboff):
Zuboff argues the true problem is surveillance capitalism, where large corporations control new technologies and operate with few limitations on data gathering and analysis for commercial purposes.
Therefore, the employment of technology is the main issue.
Algorithms are Not Neutral:
They are a product of "warm human and institutional choices" (Butcher) and underlying business models, not neutral technological phenomena.
Profit Motive (Fournier): The insatiable demand for attention and clicks, driven by deep insights from big data, tunes algorithms to maximize provider variables (e.g., time spent on site, data collected) rather than user enlightenment or democratic health.
Fournier fears a "world without mind" will be the result.
Butcher describes how algorithms have transformed news and journalism, while Gillespie highlights companies' rhetoric to maintain an "air of neutrality." Ultimately, pursuit of profit is the driver.
Critiques of Surveillance Capitalism:
Fournier focuses on monopolistic ambitions of tech companies.
Zuboff contends surveillance capitalism is a novel phenomenon that defies traditional analytical tools like monopoly and privacy, seeing users as neither consumers nor employees.
Lanier questions "who owns the future" in data-monetizing societies, problematizing distribution effects and advocating for a radical new approach to data handling.
Author's Scope: Acknowledges these broader societal issues but focuses on mechanisms threatening liberty even without phenomena like surveillance capitalism.
Approach: Non-technical, aiming to present a simplified but credible picture of:
The basic workings of big data and AI.
A slightly evolved future, positing that these issues will become more pressing.
Focus on fundamental issues, not specific implementations.
Three Foundational Propositions (Premises):
: Human nature's tendency to seek corroborating information.
: Algorithms and the creation of filter bubbles.
: Digital spaces enabling the congregation of like-minded people.
Human Nature: Selective Exposure and Confirmation Bias
Human Peculiarity: Humans exhibit biases in information seeking, are not fully rational, and vulnerable to poor decision-making (Kahneman, nudging literature).
Selective Exposure:
Individuals' propensity to seek information that aligns with their views and beliefs, while avoiding content that is different or challenging.
Originated in the 1960s but gained renewed attention with algorithmic curation.
Personalization makes avoiding uncomfortable information and seeking supporting views significantly easier than ever before.
It leads to increasingly selective exposure, even if the strength of the effect is debated.
Cognitive Dissonance (Festinger): The reason for avoiding conflicting information is the discomfort it causes; people seek to reduce exposure to aversive stimuli.
Garret's Research: Suggests personalization leads to greater exposure to supporting information, but not necessarily less exposure to conflicting information.
Confirmation Bias:
A tendency to seek or interpret information in ways that support existing beliefs, expectations, or hypotheses.
This is an unconscious process; people are unaware they are doing it.
Nickerson's analogy: The unconscious mind acts like a lawyer building a case, aiming to win rather than shed equal light on all aspects.
Empirical Evidence: Eisinger and Hahn, along with Jonas et al., show that people prefer news sources expected to agree with them, with stronger effects in real-life, sequential information settings.
Shifting Focus: While Festinger initially linked selective exposure to conscious choice, the rise of algorithms makes unconscious selective exposure of greater interest today.
Proposition One (): Individuals have a propensity to seek information that corresponds to their pre-existing conceptions, through such mechanisms as selective exposure and confirmation bias.
Algorithms, Curation of Information, and Filter Bubbles
Attention as a Scarce Resource: Filtering information has always occurred; algorithms are the primary modern tool (computer code transforming input to output).
Algorithms are Not Neutral:
They are not merely objective processes; they are designed with inherent biases and goals.
Facebook Example (Butcher): Algorithms choose what users see based on affinity, content weight, and time decay. They create clear rules for gaining attention and impose "severe sanctions" (invisibility) for violations.
Goal-Oriented Design: Social media algorithms are instructed to maximize user engagement (time on site), which in turn maximizes revenue from advertising and data collection.
Like any model, algorithms include some things and exclude others, reflecting the explicit and implicit values of their designers.
Keepers of Cultural Discussion (Gillespie): Digital platforms, driven by profit, have become primary curators of cultural discourse, raising concerns due to their unregulated operation; they are not neutral or philanthropic.
Filter Bubbles (Pariser):
People often don't understand how algorithms work.
An algorithm acts as a filter that selectively lets some information through.
When this filter is the primary source of an individual's information, it metaphorically creates a "filter bubble," skewing their perception of reality.
Customization, personalization, and social networks, driven by big data about user preferences, lead Facebook and other sites to create distinct filter bubbles for individuals.
Two Factors Making Algorithmic Curation Problematic:
Learning and Narrowing: Algorithms, initially naive, become increasingly adept at providing desired content as users make choices (likes, dislikes, time spent), leading to a narrower perceived world by filtering out upsetting content.
Undeclared Goals of Creators: Algorithms are programmed to achieve specific goals (e.g., profit) that may be undetectable to users and unrelated to their preferences. This means the world seen through algorithms is not a neutral representation, even for individuals not exhibiting selective exposure.
Growing Importance: The way algorithms shape our perception is increasingly important as customization and personalization (explicit or implicit) grow.
Social Media's Role: Social networks lead in personal information curation; about of Americans get news via social media, with Facebook being the most popular (Flaxman et al. noted that traditional media are less prone to filter bubbles as they are still early in personalization).
Proposition Two (): Algorithms powered by big data play an increasingly large part in curating the overwhelming amount of information that is available in modern society. Such algorithms are not neutral and will always choose what information to present to users based on some predetermined logic.
Digital Social Spaces: Intragroup Homogeneity and Intergroup Heterogeneity (Echo Chambers)
Historical Contact with Diverse Opinions: Historically, finding like-minded individuals in physical spaces (work, family, town) was difficult, often requiring interaction with different opinions. Options for seeking out similar views (relocation, pen pals) were limited, meaning most life interactions involved diverse perspectives.
Digital Ease: Today, it's easy to create online groups (e.g., for Hobbesians), befriend similar people, and build social networks predominantly composed of like-minded individuals.
Family and childhood friends with differing views can be easily hidden from news feeds, making them "hypothetical friends."
Algorithms further this process by prioritizing content based on affinity, even without explicit user action (Butcher, Gillespie).
New Digital World: In this curated space, individuals see news, comments, and opinions primarily from their like-minded fellows, reinforcing their views and creating an "echo chamber."
Echo Chamber: A space where individuals are largely exposed to conforming opinions, leading to increased polarization (Sunstein, Alcott, Genskow).
Impact on Diversity: A study showed that only of liberals' friends had opposite ideological views, and for conservatives.
Mass Media vs. Digital Media: Mass media could cater to binary positions (e.g., pro/con Brexit), but digital media allows catering to much smaller groups, even individuals. This process fragments large groups into smaller ones, increasing intragroup homogeneity (similarity within groups) and intergroup heterogeneity (diversity between groups).
Consequences of Homogeneity: In echo chambers, people are comfortable, their opinions are reinforced, and they lose the inclination to proactively discuss ideas with different groups. This is dangerous because groups tend to take more extreme positions than individuals.
Reduced Cross-Cutting Content: Individuals are less exposed to diverse viewpoints on social media, with friends (chosen for like-mindedness) being the primary source of news and opinions.
Compounding Effects: When combined with selective exposure (P1) and algorithmic curation (P2), the resulting "second-order diversity" (Gerken) and "enclave discussions" (Sunstein) warrant a serious warning.
Proposition Three (): In the digital world, it is easy to create areas in which people can associate with like-minded people, in contrast to our lives in physical space where, a) fewer like-minded people are accessible, and b) where there are greater chances of encountering people with various beliefs. Such areas can become what are referred to as echo chambers.
Understanding Liberty
Berlin's Facets of Liberty:
Positive Liberty: "Self-mastery," meaning the dominance of an authentic self over a merely empirical self. It involves being a subject, one's own master, and autonomous/self-directed.
Negative Liberty: The absence of interference from other men. It does not concern interference from nature or internal obstacles like desires or irrationality.
Critiques of Berlin: Callam and Skinner criticize the idea of two separate forms of liberty. Carter seeks an "overall freedom" compatible with both negative liberty and the idea that freedom includes ability.
Liberty as Non-Domination (Pettit):
This concept is positioned between positive and negative liberty.
Domination: Defined as the capacity to interfere in choices a person could otherwise make.
Non-domination requires both the absence of actual interference and security for this absence. If interference is possible at any time, one is not fully free.
Liberty and Tyranny as Political and Social Phenomena:
The author quotes: "There is a no tyranny more dangerous than an invisible and benign tyranny, one in which subjects are complicit in their victimization and in which enslavement is a product of circumstance rather than intention."
Tocqueville's Tyranny of the Majority: Concerned with the social and cultural phenomenon of liberty being limited, particularly freedom of discussion and thought, leading to a "lack of independence of mind." This can be seen as a lack of positive liberty.
Both Mill and Tocqueville stressed the need for access to a wide range of facts and opinions for freedom to be meaningful.
The author focuses on tyranny of opinion rather than the subjugation of a minority by a simple majority, highlighting Tocqueville's description of it as a "social and cultural phenomenon" in America.
Mill's View: Mill noted Tocqueville's tyranny was "not over the body but over the mind," emphasizing liberty of spirit and opinion, which is threatened or preserved by society, not just politics.
America's Unique Tyranny: Tocqueville observed an entire society acting in concert, molding individuals' natures through moral power of the majority, without needing oppressive laws.
Author's Argument: A tyranny over "hearts and minds" that strips individuality is as much a threat as physical oppression.
"Superficial Consent": Modern tyranny comes to power subtly, often with implicit consent because it's hard to perceive and arises gradually. Zuboff notes it begins with stealth and lawlessness, and informed consent is rarely given for accessing "vital social infrastructure."
Redefining Coercion: Tocqueville found classical coercion unrecognized. Horowitz used "psychic coercion" for majority uniformity. The author supports broadening coercion to include psychological dimensions (Sietra, nudging, big data).
Oppressor and Oppressed: In Tocqueville's America, every member was both, making it hard to rebel. Conformity was "freely chosen" (Horowitz); the oppressed glorified their oppression.
Modern Tyranny's Punishment: Not chains or executions, but being disregarded, ignored, shunned, ostracized, made an "alien." Being rendered "invisible" is a contemporary form of this punishment, a "life worse than death."
Freedom of Spirit and Opinion: Liberty demands not just physical freedom but also the ability to express and think contrary ideas without fear of severe social sanctions or ostracism.
Proposition Four (): Liberty requires, in addition to freedom from physical interference, an absence of social and psychological interference/domination that blocks the freedom of spirit and opinion.
Information and Individuality
Mill's "On Liberty" and Individuality:
Mill emphasized the importance of exposure to novel opinions (regardless of truth value), viewing the "quiet suppression of half of the truth as the formidable evil."
He believed truth is hardened into prejudice when people only listen to one side.
Individuality is crucial; it makes each person "more valuable to himself and is therefore capable of being more valuable to others."
It prevents values from developing in lockstep.
For Mill, interference, denial of choice, coercion, and encroachments on individuality are abhorrent (Hamburger).
Individuality is both a precondition and a guarantee of liberty.
Requirements for Individuality and Liberty:
Access to Variety: Individuality requires diverse information for inspiration and to realize other possibilities. Liberty is meaningless without options.
Informed Citizens: A functioning liberal democracy needs citizens with access to varied information to make reasonable choices.
Beyond Access: Exposure: Given human selective exposure and profit-driven algorithms, the author argues for "a certain level of minimum exposure" to diverse information, not just access.
Countering Effects: Sunstein advocates for "serendipity," while Vanderhoven explores technical solutions to combat filtering effects.
Proposition Five (): Access and exposure to a broad array of information is a requirement for the development of individuality.
The Threat of a Tyranny of Perceived Opinion: Turning Propositions into Premises
Summary of Propositions (Premises for Conclusion):
: Individuals naturally seek information supporting pre-existing beliefs and avoid conflicting information (selective exposure, confirmation bias).
: Non-neutral algorithms curate the vast amount of information in modern society, based on predetermined logic.
: Digital technology and social networks facilitate and encourage the congregation of like-minded people, minimizing exposure to conflicting opinions (echo chambers).
: Liberty extends beyond freedom from physical interference to include freedom from social and psychological interference/domination that obstructs freedom of spirit and opinion.
: Access and exposure to a broad range of information are essential for the development of individuality.
First Conclusion (from , , ): These three premises combined create a strong potential for individuals to live lives predominantly exposed to information that supports pre-existing opinions, minimizing contact with opposing views.
Algorithms narrow views, both unintentionally and for creator goals.
Human tendencies confirm preconceptions, shrinking one's "window to the world to a slit."
Congregating with like-minded people provides social support but further narrows perspectives.
This leads to radical differences in perceived reality between groups (high intergroup heterogeneity) and similarity within groups (high intragroup homogeneity).
This effect is termed the "compound confirmation effect of information technology and social networks."
Second Conclusion (from , , , and ): The compound confirmation effect, by making access to varied information difficult, poses a direct threat to individuality. If individuality is compromised or lost, the concept of liberty becomes largely meaningless.
Main Conclusion: When all five premises are considered together, they depict a substantial threat to liberty, manifesting as a "tyranny of perceived opinion."
How Predictions and Expectancies Threaten Liberty
The Two-Headed Creature of Tyranny:
Human Head (Tocqueville's Resurrection): A tyranny where the moral power of popular opinion is strong and reinforced by group homogeneity.
Online Behavior: When encountering displeasing content from a friend, options are: a) ignore (hard), b) argue (costly, public risk), c) hide (attractive, problem disappears).
Modern "Punishment": Hiding a person online mirrors Tocqueville's tyrant's punishment of dissenters – being shunned, ignored, put out of sight, and denied a dignified response. This possibility fosters fear of invisibility (Butcher) and shapes what we say and proclaim.
Self-Imposed Conformity: The strong human desire for consistency, combined with the pressure to conform (e.g., half-heartedly liking content), unconsciously locks individuals into the "cage of opinions" of their group.
Blackford's "Cyber Mobs": While the author doesn't focus on active online moral policing or shaming, these are acknowledged as potential amplifications of tyranny.
Machine Head (Predictions and Expectancies):
Algorithms and big data subtly determine individual development, limit alternatives, and hinder resistance to presented paths (Andrejevic, Cukier).
Self-Fulfilling Prophecies: Big data predictions of preferences can lead people to conform to those predictions, creating self-fulfilling prophecies (Zuboff's "behavioral surplus").
Evaluating Accuracy: If predictions shift behavior, their accuracy becomes questionable (Lanier).
Unpreparedness: Society is ill-prepared to assess these ethical issues, with unintended consequences often resulting from new technology deployment (Zwitter).
"Shackling Users to Profiles": Dwork and Mulligan argue big data classification "shackles" users to profiles (the "dividual") that determine relevant information, potentially based on incomplete or approximated data.
This section focuses on three psychological/sociological theories explaining how expectations shape reality:
The Expectancy Effect (Behavioral Confirmation Theory): My expectations of someone influence how they act and perform.
Individualized profiles on websites and services create environments users expect to be comfortable with, making these expectations self-fulfilling.
Suggestions (e.g., on Facebook, Netflix) are perceived as objective but are partly based on user history, similar users' behavior, and the provider's desired consumption. Algorithm non-neutrality is key.
The Proteus Effect: People change their behavior based on how they perceive themselves, even if the change is in a digital setting and unseen by others.
Based on Self-Perception Theory, people infer attitudes by observing their own behavior (e.g., avatar changes altering real-world behavior).
Examples: Attractive avatar -> closer to others, more disclosure. Tall avatar -> more confidence. Sexualized portrayals -> self-objectification, acceptance of rape myths.
Social media portrayal frameworks (e.g., Facebook highlighting certain posts) can affect changes in user behavior and feelings of self.
The Thomas Theorem: "If people define situations as real, the consequences of these situations are real." (Self-fulfilling prophecies).
Believing a prediction increases its likelihood of coming true (e.g., banking crises, failing a test).
Algorithmic expectancies can be perceived as prophecies, which users then make reality.
Big data's inductive nature (based on historical facts) creates a conservative force, reinforcing old definitions of reality even if alternatives were possible (Bins on fairness in algorithms).
Merton's example of ethnic conflict shows how self-generated "facts" perpetuate issues. These predictions have simultaneous internal and external effects on liberty.
Amplification: The level of personalization of expectations and representations made possible by big data and algorithms significantly strengthens these effects.
Conclusion
Alexis de Tocqueville's fears of an invisible, immaterial, benign, and voluntary tyranny of public opinion are more relevant than ever.
Such a tyranny, though hard to perceive, is as dangerous to liberty as older forms.
The author argues that a tyranny over minds (not bodies), consented to due to lack of perception, poses an even greater danger.
The modern threat stems from the "compound exposure effect" created by three factors:
Our tendency to seek corroborating facts ().
The use of non-neutral algorithms as information curators ().
The ease of creating digitally exclusive networks with like-minded individuals ().
Within these echo chambers and filter bubbles, the tyranny of perceived opinion is a clear and present danger to liberty.
Impact on Liberty:
Negative Liberty: While perhaps not most threatened directly, the absence of social/psychological interference () is vital.
Positive Liberty: Requires a form of autonomy that demands access to the necessary information, which is compromised by these phenomena.
Liberty as Non-Domination (Pettit): Algorithms and their creators clearly possess the capacity to arbitrarily interfere () with our choices regarding understanding the world and making life decisions.
The author deliberately focused on intrinsic threats to liberty, not just explicit manipulation or abuse, although including those aspects (referencing Sietra) would significantly strengthen the argument.
Societal Implications: A liberal, healthy society requires well-informed citizens with a sufficiently similar worldview to enable collective discussion, debate, and deliberation.
The decreased breadth of information has clear implications for individuality and liberty ().
The author stresses the paramount importance of addressing these issues in political science and law, rather than leaving them exclusively to communication studies and information technology, advocating for the inclusion of political philosophy (Kockelberg) in this critical debate.
Introduction
The digital age, despite offering vast information access, creates challenges for effective utilization due to sophisticated curation technologies like algorithms and AI. This leads to filter bubbles and echo chambers, reinforced by selective exposure, where individuals primarily encounter information aligning with their beliefs, avoiding contradictory content. This curation becomes leading, threatening liberty by undermining individuality and autonomy. The author likens this to a "tyranny of algorithms and perceived opinion," echoing Tocqueville's fear of the "tyranny of the majority." The article highlights how information abundance can paradoxically reduce exposure to diverse ideas, hindering the development of individuality and liberal societies.
Background and Terminology: Big Data Information
The current era is defined by Big Data (Volume, Velocity, Variety), synergizing with Artificial Intelligence (AI) and Machine Learning (ML). The central issue is not the technology itself but its employment, particularly surveillance capitalism (Zuboff), where corporations use new technologies for data gathering and analysis with few limitations. Algorithms are not neutral; they are products of "human and institutional choices" and profit motives, tuning algorithms to maximize engagement rather than user enlightenment. This creates problems like monopolistic ambitions (Fournier) and new forms of data monetization (Lanier).
Three Foundational Propositions (Premises):
: Human nature tends to seek corroborating information.
: Algorithms create filter bubbles.
: Digital spaces enable the congregation of like-minded people.
Human Nature: Selective Exposure and Confirmation Bias
Humans exhibit biases, are not fully rational, and are prone to seeking information that confirms their existing views (selective exposure) while avoiding conflicting content due to cognitive dissonance. This process is often unconscious (confirmation bias), where the mind interprets information to support existing beliefs. Algorithmic curation exacerbates this, making it easier to avoid uncomfortable information. : Individuals have a propensity to seek information that corresponds to their pre-existing conceptions, through such mechanisms as selective exposure and confirmation bias.
Algorithms, Curation of Information, and Filter Bubbles
Algorithms are the modern tools for information filtering. They are not neutral but are designed with inherent biases and goals, often to maximize user engagement and profit (e.g., Facebook's affinity-based content selection). These platforms become "keepers of cultural discussion." Filter bubbles (Pariser) arise when algorithms, acting as filters, customize information based on user data, narrowing individuals' perceptions of reality. Problems intensify as algorithms learn to provide desired content and due to the creators' undeclared, profit-driven goals. : Algorithms powered by big data play an increasingly large part in curating the overwhelming amount of information that is available in modern society. Such algorithms are not neutral and will always choose what information to present to users based on some predetermined logic.
Digital Social Spaces: Intragroup Homogeneity and Intergroup Heterogeneity (Echo Chambers)
Unlike physical spaces where diverse interactions were common, digital platforms facilitate easy congregation of like-minded individuals, creating echo chambers. These spaces reinforce views, leading to increased intragroup homogeneity (similarity within groups) and intergroup heterogeneity (diversity between groups). This reduces exposure to diverse viewpoints and can lead groups to adopt more extreme positions. : In the digital world, it is easy to create areas in which people can associate with like-minded people, in contrast to our lives in physical space where, a) fewer like-minded people are accessible, and b) where there are greater chances of encountering people with various beliefs. Such areas can become what are referred to as echo chambers.
Understanding Liberty
Liberty is considered broadly, encompassing more than just freedom from physical interference. It includes positive liberty (self-mastery) and non-domination (Pettit), which is the absence of the capacity to interfere. The author focuses on tyranny of opinion (Tocqueville, Mill), a social and cultural phenomenon that limits freedom of discussion and thought by molding individuals through the moral power of the majority. Modern "punishments" involve being ignored, shunned, or made "invisible" online, fostering conformity. : Liberty requires, in addition to freedom from physical interference, an absence of social and psychological interference/domination that blocks the freedom of spirit and opinion.
Information and Individuality
Mill's "On Liberty" stresses the importance of exposure to novel opinions for the development of individuality, which is both a precondition and a guarantee of liberty. Individuality requires diverse information for inspiration and informed choice. The author argues for "a certain level of minimum exposure" to varied information, not just access, given human selective exposure and profit-driven algorithms. : Access and exposure to a broad array of information is a requirement for the development of individuality.
The Threat of a Tyranny of Perceived Opinion: Turning Propositions into Premises
Combining , , and creates a "compound confirmation effect": individuals are exposed predominantly to information supporting their existing views, narrowing their "window to the world." This leads to high intergroup heterogeneity and intragroup homogeneity. When combined with (necessity of diverse info for individuality), this effect directly threatens individuality, making liberty meaningless. The culmination of all five propositions reveals a substantial threat to liberty, manifesting as a "tyranny of perceived opinion."
How Predictions and Expectancies Threaten Liberty
This modern tyranny has two heads: Human (Tocqueville's fear of moral power of popular opinion, reinforced by online shunning/invisibility and self-imposed conformity) and Machine (algorithms subtly determine individual development, limit alternatives through predictions). This section explains how expectations shape reality through three psychological theories:
The Expectancy Effect (Behavioral Confirmation Theory): Expectations influence behavior; algorithmic suggestions can create self-fulfilling prophecies.
The Proteus Effect: Digital self-perception (e.g., avatar) alters real-world behavior.
The Thomas Theorem: "If people define situations as real, the consequences of these situations are real;" algorithmic predictions, perceived as prophecies, can become reality. These effects are significantly amplified by big data and algorithm personalization.
Conclusion
The author concludes that Tocqueville's