{"CACHEDAT":"2026-08-19 10:12:25","DEEPLQUOTAEXCEEDED":false,"TRANSLATEDAT":"","SOURCESIGNATURE":"15475565b52f9545c2e6dffdfdc4967544547ff49c84ace5d0e84630a63f3664","SLUG":"information-visibility-in-digital-information-environments-1d4XJXs8gX","MARKDOWN":"# Information Visibility & Prominence {{visible_information prominence}}\n\nInformation visibility refers to whether — and how prominently — specific information items are encountered by users in digital environments. Information that exists in a platform's index or network does not automatically reach all users, nor reach them in the same way. \n\n\n:::info\nWhat users actually see is rarely the product of a single mechanism.\n\n* In a search engine, an algorithmically ordered list of organic results is presented alongside paid placements, AI-generated summaries, and sometimes editorial highlights — and the underlying ranking signals can be deliberately influenced through Search Engine Optimisation.\n* In a social media feed, algorithmically ranked posts appear next to sponsored content, recommended accounts, and trending sections. Each component follows its own logic and contributes to a composite visibility outcome.\n\n:::\n\nVisibility is shaped by several interacting processes:\n\n* **user activity: environment selection, engagement, circulation, re-circulation, re-creation**\n* **gatekeeping: inclusion & exclusion**\n* **source-driven self-promotion**\n* **platform-side curation**\n\nFrom the interplay of these processes, two opposite dynamics emerge:\n\n* **information amplification**: visibility is broadened across users\n * Amplification results from the above processes and feeds back into them: once an information item has gained visibility, the resulting engagement signals feed back into platform-side curation and make the information item more likely to be surfaced to further users (→ algorithmic amplification). \\nThis dynamic can also be triggered deliberately by coordinated actors manufacturing engagement (→ coordinated amplification).\n* **information narrowing**: the range of perspectives reaching a user is restricted\n * Narrowing is self-reinforcing too, but it restricts rather than broadens: algorithmic personalisation keeps surfacing similar information items (→ filter bubble), while a user's own choice of sources and contacts narrows the range of perspectives they meet (→ echo chamber).\n\n\n:::warning\n**Distinguishing reliability and visibility is essential for information literacy.**\n\n* **Reliability** depends on who created an information item (author) and how it was edited (editorial review), not on whoever shared it, on the information environment it appeared in, or on how prominently it was surfaced.\n* **Visibility** depends on who shares an information item, on the mechanisms that surface it, and on the practices through which it is amplified. Visibility is not a measure of reliability.\n\n:::\n\n# Visibility Phenomena {{visibility phenomenon}}\n\nThe phenomena below are the concrete outcomes that decide what an individual user actually encounters. \n\n| Phenomenon | Visibility Effect | Trigger | Mechanism | Dynamic |\n|------------|-------------------|----------|-----------|---------|\n| **Echo chamber** | A user is mainly exposed to reinforcing views | Environment selection | Social reinforcement | Information narrowing |\n| **Filter bubble** | A user is increasingly shown content matching their algorithmic profile | Engagement | Algorithmic personalisation | Information narrowing |\n| **Trending** | A user is likely to see a topic or hashtag because the platform flags it as trending and gives it a dedicated slot | Circulation | Trending detection | Information amplification |\n| **Virality** | A user is likely to see a single item because so many accounts pass it on that it keeps reappearing | Re-circulation | Algorithmic amplification | Information amplification |\n| **Spill-over effect** | A user is likely to encounter some information because it occurs in different information items across more than one information environment | Re-creation | Cross-environment diffusion | Information amplification |\n\nFilter bubbles, trending, and virality start from a → user activity (the trigger) and are produced by mechanisms that are platform-side and **beyond the user's control**. \n\nEcho chambers also start from a user activity, environment selection, but their mechanism is social reinforcement within a self-chosen group, not a platform-side one. \n\nThe spill-over effect differs on both counts: it is triggered and carried by actors, often journalists or editors, rather than by the user who later encounters the information.\n\n\n:::info\nEach phenomenon makes a user either more likely to encounter a particular information item or topic (→ information amplification) or less likely to encounter diverging perspectives (→ information narrowing).\n\n:::\n\n# User Activity {{interaction_user interaction_activities}} \n\nUser activity covers everything a user or actor does that bears on what becomes visible. Five types matter here, each the trigger of one of the phenomena above:\n\n* **environment selection**: choosing which environments to use and whom to follow (→ echo chamber)\n* **engagement**: interacting with information items already on display (→ filter bubble)\n* **circulation**: bringing a new information item into an environment by posting it (→ trending)\n* **re-circulation**: distributing existing information items within an environment by sharing, forwarding, or reposting (→ virality)\n* **re-creation**: turning an existing information item into a new, derived one by quoting, editing, or reframing it, often in another environment (→ spill-over)\n\n\n:::warning\nIn each case the user sets the process in motion, but the mechanism that follows is largely outside their control. The user's own influence on what they see is smaller than it appears.\n\n:::\n\n### Actors {{actor}}\n\nAn actor is any entity involved in handling information, independent of any particular information environment: a person, an organisation, or automated software (a bot or an algorithm).\n\n### Users {{user}} \n\nA user is any actor that acts within information environments, as a content creator, a content distributor, or a content recipient. A user need not be a person; it can also be an organisation or automated software (a bot).\n\nWhere an environment uses accounts, a user acts through one. A single user may operate several, even within the same environment (→ sockpuppet).\n\n### Bots {{bot}} \n\nA bot is a computer programme that automatically performs tasks, often repetitive ones. Bots range from\n\n* simple, harmless tools (such as web crawlers that index pages for search engines, automated testing systems, or chatbots that answer routine customer questions) to \n* malicious programmes designed to spread spam, malware, or disinformation. \n\n\n:::info\nSocial bots simulate humans in digital information environments.\n\n:::\n\n## Environment Selection\n\nEnvironment selection is the user's own choice of which information environments to use and whom to follow within them. Among the five user activities, it is the one the user controls most directly: no platform mechanism sits between the choice and its outcome. The other activities feed rocesses the user cannot control. \n\n\n:::warning\nBecause the choice belongs to the user, it is also the **most effective point of intervention**: \\ndeliberately seeking out varied information environments and sources is the clearest way \n\n* to widen one’s perspectives, and \n* to counter one's own confirmation bias, and \n* to keep out of an echo chamber.\n\n:::\n\n### Echo Chamber {{echo chambers}}\n\nAn echo chamber is a social structure in which a user is primarily exposed to opinions, claims, or viewpoints that reinforce their existing beliefs, while dissenting viewpoints are absent, dismissed, or actively discredited. \n\nCass Sunstein (2018) describes the political consequences: when groups insulate themselves from outside perspectives, internal beliefs intensify and become more extreme over time (group polarisation).\n\nAn echo chamber results primarily from → environment selection: the user's own choices about which environments to use and whom to follow, a process also called self-selection. It is therefore distinct from the → filter bubble, which is built by algorithmic personalisation on the platform's side. \n\n#### Confirmation Bias\n\nEnvironment selection is partly driven by confirmation bias, the cognitive tendency to seek out and trust information and sources that align with existing beliefs. The reinforcing effect comes from the social structure itself, not from invisible algorithmic filtering.\n\n#### Epistemic Bubble {{epistemic bubbles}}\n\nC. Thi Nguyen (2020) draws a conceptual distinction that matters for intervention:\n\n* An **epistemic bubble** is a social structure in which other relevant voices are simply absent. Its inhabitants do not hear opposing viewpoints, but they do not actively reject them.\n* An **echo chamber** in the strict sense is a social structure in which other relevant voices are actively discredited. Members may hear opposing perspectives but learn to distrust their sources.\n\nAn epistemic bubble can be opened by introducing new information. An echo chamber resists correction even when external evidence is presented, because the sources of that evidence have already been delegitimised. \n\n\n:::warning\nWidening one's environment selection can dissolve an epistemic bubble, but a fully formed echo chamber needs the harder work of rebuilding trust in excluded sources.\n\n:::\n\n\n:::warning\nEmpirical work suggests that strong, ideologically isolated echo chambers are less common than popular discourse implies (Cinelli et al., 2021; Guess et al., 2018), but where they exist, they can be highly resistant to correction. Mere agreement within a group is not in itself an echo chamber. The defining feature is the active exclusion or discrediting of outside perspectives.\n\n:::\n\n\n:::info\n* Sunstein, C. R. (2018). *#Republic: Divided Democracy in the Age of Social Media*. Princeton University Press.\n* Nguyen, C. T. (2020). Echo chambers and epistemic bubbles. *Episteme*, 17(2), 141–161. \n* Cinelli, M., De Francisci Morales, G., Galeazzi, A., Quattrociocchi, W., & Starnini, M. (2021). The echo chamber effect on social media. *PNAS*, 118(9), e2023301118. \n* Guess, A., Lyons, B., Nyhan, B., & Reifler, J. (2018). *Avoiding the Echo Chamber about Echo Chambers: Why Selective Exposure to Like-minded Political News Is Less Prevalent Than You Think*. Knight Foundation White Paper.\n\n:::\n\n### Account {{accounts}}\n\nAn account is a user's identity within a digital information environment, the profile through which they post, share, and interact. Behind it stands a person, an organisation, or automated software (a bot). Traditionally, the source of an information item was a named author, identifiable as its originator. \n\n\n:::warning\nIn digital environments, actors operate through accounts that are often pseudonymous: the account is visible, but the real identity behind it need not be disclosed, and the two do not necessarily coincide. This makes → source credibility far harder to assess: without a named author, there is often no track record, expertise, or affiliation to weigh.\n\n:::\n\n\n:::info\nThe following sections often use user rather than account for readability, although account is the more precise term, since a single user may operate several accounts within the same information environment.\n\n:::\n\n#### Username {{usernames}}\n\nA username is a unique account identifier chosen by the user when registering on a platform; no two accounts on the same platform can share one. Depending on the platform it works as a login credential, a public profile identifier, or both.\n\n#### Handle {{handles}}\n\nA handle is the public form of a username on a social media platform: the identifier under which an account can be found, addressed, and mentioned, usually shown with an @ symbol and appearing in the profile URL. On most platforms, the handle and the username are identical; a few, such as Facebook, keep a separate username for the URL and show a non-unique display name publicly.\n\n#### Display Name {{display names}}\n\nA display name is the visible label shown on a profile and beside posts. It is freely chosen, may contain spaces and special characters, and need not be unique, so many accounts can carry the same one. It cannot be used to identify or address an account.\n\n#### User ID {{user ids_userid_userids}}\n\nA user ID is the internal identifier a platform assigns to an account, typically numeric. It is not chosen by the user and normally stays fixed even when the username or display name changes. Platforms use it internally rather than for public addressing.\n\n#### Mention / @Mention {{mentinmentions_@mention_@mentions}}\n\nA mention is the act of referring to an account by typing @ followed by its handle in a post, comment, caption, or message. It usually notifies the account and links to its profile. The @ string is the handle; using it is the mention.\n\n### Account Reach {{reach_reaches_reached}}\n\nSeveral factors determine the number of users or recipients an account can potentially reach when posting or sharing an information item:\n\n* #### follower or subscriber count\n* #### verification status \n* #### account standing: age, engagement history, platform reputation\n\nReach affects visibility in two ways:\n\n* **directly**: information items shared by high-reach accounts appear in more feeds at the moment of sharing\n* **indirectly**: high-reach accounts generate more engagement-derived signals, which platform algorithms then use to elevate information items in ranking (→ algorithmic curation)\n\nReach varies widely: a private account with 100 followers and a public account with one million followers operate at fundamentally different scales of influence on visibility.\n\n## Engagement {{user action_user actions}}\n\nEngagement covers all user actions, i.e. everything a user deliberately does in an information environment, from taking content in, through responding to it, to producing their own. Adapting the COBRA model (Muntinga, Moorman & Smit, 2011), these actions fall into three ascending levels: consumption, contribution, and creation.\n\n| Consumption | \\- searching- selecting / clicking- watching / listening / reading- purchasing / downloading- saving / bookmarking |\n|-------------|-------------------------------------------------------------------------------------------------------------------|\n| Interaction | \\- liking / reacting
- following / subscribing
- rating / reviewing- re-circulating |\n| Creation | \\- creating: circulating- recreating: replying / commenting |\n\n\n:::info\nMuntinga, D. G., Moorman, M., & Smit, E. G. (2011). Introducing COBRAs. International Journal of Advertising, 30(1), 13–46. \n\n:::\n\n### Filter Bubble {{filter bubbles}}\n\nA filter bubble is a state of informational isolation produced by algorithmic personalisation, in which a user is increasingly exposed to content that aligns with their inferred preferences and past engagement, while content that diverges may be filtered out, typically without the user's awareness.\n\nThe term was coined by Eli Pariser (2011) to describe how personalisation algorithms on Google, Facebook, and similar platforms can produce systematic exposure asymmetries based on user signals. \n\n\n:::warning\nEmpirical research has substantially qualified Pariser's original thesis. Studies have found that algorithmic personalisation does shape what users see, but most users still encounter ideologically diverse content: partly because their own social networks include varied views, and partly because algorithms do not isolate as completely as the popular discourse suggests (Bakshy et al., 2015; Flaxman et al., 2016; Bruns, 2019). \n\nThe filter-bubble effect is real but typically weaker than commonly assumed; pre-internet selective exposure (e.g., choosing newspapers or TV channels) was in many cases stronger.\n\n:::\n\n\n:::info\n* Pariser, E. (2011). *The Filter Bubble: What the Internet Is Hiding from You*. Penguin Press.\n* Bakshy et al (2015)\n* Flaxman, S., Goel, S., & Rao, J. M. (2016). Filter bubbles, echo chambers, and online news consumption. *Public Opinion Quarterly*, 80(S1), 298–320. \n* Bruns, A. (2019). *Are Filter Bubbles Real?* Polity Press.\n\n:::\n\n## Circulation: Posting {{circulation_post_posts_posting_posted}}\n\nOnce an information item has been created, posting brings it into an information environment for the first time: its creator posts, sends, publishes, or broadcasts it. This initial publication is distinct from re-circulation, where an existing item is passed on by others.\n\n### Trending {{trending_trend_trends}}\n\nTrending is a platform-assigned status that indicates that a topic or hashtag is attracting a sudden surge of user activity within a short time, across the many separate posts that address it or carry it. What counts is the rate of increase, not the overall volume.\n\nTrending can emerge organically from many independent contributions, or be manufactured through coordinated posting and bot networks (→ coordinated amplification).\n\nThe platform itself also shapes it, by promoting, filtering, or suppressing what surfaces (→ platform-side curation).\n\n\n:::warning\nContent that triggers rapid reactions, such as divisive or emotionally charged topics, is more likely to trend, though whether it surfaces also depends on the platform's ranking and moderation mechanisms.\n\n:::\n\n\n:::info\n#### Hashtag {{hashtags}}\n\nA hashtag is a keyword or phrase marked with a # symbol and written without spaces. It links a post to every other post carrying the same label, turning scattered contributions into a searchable group around a shared topic.\n\nAnyone can attach a hashtag to a post, which makes it a tool for visibility: a widely-used or trending label places the post within a larger stream and exposes it to users following that label. Because a hashtag aggregates many separate posts, it can itself attract a surge of activity and trend (→ trending), which also makes it a frequent target of coordinated campaigns that push a label through mass posting (→ coordinated amplification).\n\n:::\n\n## Re-Circulation: Sharing, Forwarding, Reposting {{re-circulation_recirculation_recirculate_recirculates_recirculated_share_forward_repost_shares_forwards_reposts_shared_forwarded_reposted}}\n\nAccounts pass existing information items on by sharing, forwarding, linking, reposting, or otherwise recirculating them. Each account is operated by a person, an organisation, or automated software (a bot). \n\n\n:::warning\nWhen an information item is passed on with added commentary, interpretation, or reframing, a new information item is created with a new author (→ re-creation).\n\n:::\n\n### Virality {{viral}}\n\nVirality is the phenomenon by which a single information item spreads rapidly through re-circulation, much as a virus propagates: users share or repost it, others pass it on in turn, and the resulting engagement raises its ranking, surfacing it to still more accounts (→ algorithmic amplification).\n\nAn item is more likely to be shared when it evokes high-arousal emotions, moral reactions, or out-group animosity, especially in political or conflict-oriented contexts. High → account reach at the origin is not required: small accounts can also produce highly viral items.\n\nVirality can emerge organically or be manufactured through coordinated sharing and engagement (→ coordinated amplification).\n\n#### ☑ Virality vs. Trending\n\n| Feature | Virality | Trending |\n|---------|----------|----------|\n| **What gains visibility** | A single information item, such as a specific post, video, or image | A topic or hashtag: not one specific item, but the many separate posts that address the same topic or carry the same hashtag |\n| **Primary drivers** | Users share, repost, or forward the item to others, who pass it along in turn; this cascading spread can be reinforced when the resulting engagement raises the item's ranking (→ algorithmic amplification) | Many separate posts address the same topic or carry the same hashtag within a short time; the platform reads how fast this activity is rising, not how much there is in total, and highlights the topic or hashtag as a whole (→ trending detection) |\n| **Time pattern** | Can continue over longer periods | Time-bound; lasts as long as the rate of activity stays high or the platform keeps surfacing it |\n| **How it can be manipulated** | Coordinated sharing, bot amplification, or artificial engagement directed at the specific information item (→ coordinated amplification) | Coordinated posting campaigns, or manufactured trends through bot networks (→ coordinated amplification) |\n\n\n:::info\nBoth virality and trending can emerge organically, be pushed by coordinated actors (→ coordinated amplification), or be shaped by the platform's own curation. Both also tend to favour emotionally arousing, morally charged, or divisive content, especially in political or conflict-oriented contexts.\n\n:::\n\n\n:::info\n* **Berger, J., & Milkman, K. L. (2012).** What makes online content viral? *Journal of Marketing Research*, 49(2), 192–205. \n* **Lee, J., & Umback, J. (2026).** The viral turn: rethinking virality in the creator economy on TikTok. *Continuum*, 1–26. \n* **Maarouf, A., Pröllochs, N., & Feuerriegel, S. (2024).** The virality of hate speech on social media. *Proceedings of the ACM on Human-Computer Interaction*, 8 (CSCW1), 1–22. \n* **Rathje, S., Van Bavel, J. J., & van der Linden, S. (2021).** Out-group animosity drives engagement on social media. *Proceedings of the National Academy of Sciences*, 118(26), e2024292118. \n* **Rathje, S., & Van Bavel, J. J. (2025).** The psychology of virality. *Trends in Cognitive Sciences*, 29(10), 914–927. \n* **Rogers, E. M. (2003).** *Diffusion of Innovations* (5th ed.). New York: Free Press.\n* **Sangiorgio, E., Cinelli, M., Cerqueti, R., & Quattrociocchi, W. (2024).** Followers do not dictate the virality of news outlets on social media. *PNAS Nexus*, 3(7), pgae257. \n* **Schlessinger, J., Garimella, K., Jakesch, M., & Eckles, D. (2023).** Effects of Algorithmic Trend Promotion: Evidence from Coordinated Campaigns in Twitter's Trending Topics. *Proceedings of the International AAAI Conference on Web and Social Media (ICWSM)*, 17(1), 777–786. \n* Asur, S., Huberman, B. A., Szabo, G., & Wang, C. (2011). *Trends in social media: Persistence and decay*. arXiv. \n* Herman, L. M. (2023). *For who page? TikTok creators’ algorithmic dependencies*. Proceedings of the IASDR Conference. \n* Jenkins, H., Ford, S., & Green, J. (2013). *Spreadable media: Creating value and meaning in a networked culture*. New York University Press.\n* Lee, J., & Umback, J. (2026). The viral turn: Rethinking virality in the creator economy on TikTok. *Continuum, 40* (2), 319–344. \n* Maarouf, A., Pröllochs, N., & Feuerriegel, S. (2024). The virality of hate speech on social media. *Proceedings of the ACM on Human-Computer Interaction, 8* (CSCW1), 1–22. \n* Rathje, S., & Van Bavel, J. J. (2025). The psychology of virality. *Trends in Cognitive Sciences, 29* (10), 914–928. \n* Rathje, S., Van Bavel, J. J., & Van der Linden, S. (2021). Out-group animosity drives engagement on social media. *Proceedings of the National Academy of Sciences, 118* (26), 1–9. \n\n \n\n:::\n\n## Re-Creation: Quoting, Editing, Remixing {{re-creation_recreation_re-create_re-creates_re-created_re-creating_quotes_edits_remixes_quoted_edited_remixed}}\n\nA user or actor re-creates an existing information item as a new, derived one by quoting, commenting on, editing, summarising, remixing, reframing, or embedding it in a new context. Unlike re-circulation, which passes the same item on unchanged, re-creation produces a new information item with a new author. That author becomes a secondary source: the derived item carries their selection and framing, not only the original content.\n\n### Spill-Over Effect {{diffusion_cross-environment diffusion_spill-over_spillover_spills over_spilled over}}\n\nThe spill-over effect is the phenomenon by which the same information comes to appear across more than one information environment, reaching audiences beyond its origin. It arises through **cross-environment diffusion**, when re-creation carries information across a boundary: an actor turns the content of one information item into a secondary item in another environment, through journalistic reporting, editorial curation, or a user quoting it elsewhere.\n\n* A research finding shared on a scholarly forum may be discussed on social media and then summarised by an AI assistant.\n* A topic artificially amplified by bots on a social media platform may be picked up by journalists and reported as news.\n\n\n:::info\nDiffusion increases the reach of both reliable and unreliable information. Because the same information then appears in several environments at once, it also tends to look more widely established than it is.\n\n:::\n\nDiffusion can occur organically, or be set in motion by manipulation to push content into new environments (→ coordinated amplification).\n\n#### Epistemic Laundering\n\nEpistemic laundering is a specific effect of cross-environment diffusion. It occurs when the secondary source is considered more credible than the original one, for example when a claim from an anonymous post reappears in a formal publication or a peer-reviewed paper. The information then tends to seem more reliable purely through having moved, even though the underlying claims and evidence are unchanged. \n\n\n:::warning\nEpistemic laundering reveals a tendency in recipients: because a source seems credible, they assume the information it carries is reliable.\n\n:::\n\n\n:::success\nStokel-Walker, C. (2026). Scientists invented a fake disease. AI told people it was real. Nature, 652(8110), 559-561. \\nA team at the University of Gothenburg, led by a medical researcher, invented a fake skin condition called Bixonimania to test whether AI systems would absorb and repeat medical misinformation. They presented it as a supposed condition linked to blue-light exposure from screens, with symptoms such as sore, itchy eyes and a pinkish hue on the eyelids. They then created deliberately fake academic-looking preprints, planted with obvious warning signs: a fictional author with an AI-generated photo, a non-existent university, and references to Starfleet Academy and the USS Enterprise. \n\nWithin weeks, major AI chatbots began reproducing Bixonimania as a real medical condition, in some cases offering users explanatory or health-related advice. In parallel, the fake material was cited in at least one published paper, since retracted, in the Springer Nature journal *Cureus*. Nature reported that the preprints have since been removed from [Preprints.org](http://Preprints.org).\n\nCross-environment diffusion: log posts → fake preprint → webcrawlers → AI chatbot answers → academic citation\n\n:::\n\n# Gatekeeping {{gate-keeping}}\n\nIn its broadest definition, gatekeeping is the overarching process of controlling which information is selected and framed into information items that reach people (Shoemaker & Vos, 2009). In this sense, it covers both the formation of an information item and its passage to an audience.\n\nHere, gatekeeping is used in a narrower sense: the control over which existing information items are admitted to or excluded from visibility. As such, it is distinguished from source-driven self-promotion and platform-side curation: gatekeeping decides whether an information item is visible at all, curation and promotion how prominently a visible information item is presented.\n\nTraditionally, decisions over information visibility were made by humans: editors (→ editorial review). In the course of digital transformation, they have become increasingly complemented by automated or semi-automated algorithmic processes (→ crawling & indexing, → content moderation).\n\nGatekeeping operates at three points in an information item's life: \n\n* before publication: \n * editorial review\n* after publication: \n * inclusion: crawling & indexing \n * exclusion: filtering & content moderation\n\n## Pre-Publication Editorial Review\n\nFor many decades, editorial review was the standard process in information environments with a traditional editorial structure, where journalists, editors, and publishers decided which items were published and which were not ==(Shoemaker & Vos, 2009)==. Where no editorial structure exists, this pre-publication gate is absent: anyone can publish without review, and gatekeeping shifts to the post-publication processes below.\n\n* information environments with an editorial structure:\n * print media, broadcast media\n * digital platforms of print and broadcast media and digital publishers\n* actors:\n * humans: journalists, editors, publishers\n\nIn digital media, the human function is increasingly complemented by algorithmic selection; research describes the result as a hybrid coexistence rather than a wholesale replacement (Wallace, 2018).\n\n## Post-Publication Gatekeeping\n\nAfter publication, gatekeeping operates in two ordered steps: \n\n* **inclusion:** which information items are captured by information access systems \n* **exclusion:** which information items are removed\n\n### Inclusion: Crawling & Indexing {{crawl}}\n\nInformation access systems crawl the web and decide which existing information items they capture and make available, either for display (→ search systems, → discovery systems) or for processing (→ generative AI systems).\n\n* information access systems\n * search systems: index crawled information items for retrieval\n * discovery systems: aggregate and catalogue information items into feeds\n * generative AI systems: ingest information items for synthesis\n* actors\n * mostly algorithmic, with human intervention at selective points (source selection)\n\n\n:::info\nIn digital information channels & platforms, where users post directly, the inclusion gate is mostly open, unless there are access restrictions.\n\n:::\n\n### Exclusion: Filtering & Content Moderation\n\nBecause the inclusion gate is open in digital information channels & platforms, where users post directly, filtering and content moderation happen after publication in a second step. Information items in the platform are checked against platform policies and, if necessary, removed or blocked. Information access systems apply the same exclusion to their index.\n\n* digital information channels & platforms, where users post directly, and information access systems\n * SPAM filters: remove unsolicited or automated bulk content\n * content filters: remove policy-violating content (hate speech, illegal content, graphic material)\n* actors\n * hybrid: human moderators (fact-checkers and/or users) and algorithmic filters\n\nUnlike pre-publication editorial review, which combines gatekeeping with systematic quality control and source verification, post-publication moderation has only a weak quality-control component (→ fact-checking): some platforms flag or label false content, but this function is limited and, on major social media platforms, in retreat. \n\n\n:::info\nMeta ended its third-party fact-checking programme in January 2025, replacing it with user-driven community notes: .\n\n:::\n\n# Source-Driven Self-Promotion\n\nSources — content creators, publishers, advertisers, and website operators — actively promote their information items to reach audiences. \n\nUnlike platform-driven mechanisms, which shape visibility algorithmically, sources apply the following strategies to influence how prominently their content appears:\n\n* targeting specific recipients or accounts (→ direct addressing), \n* adapting content to platform ranking systems (→ SEO, SMO, platform-specific optimisation),\n* paying directly for visibility (→ influencer partnerships, paid placements).\n\n## Direct Addressing\n\nDirect addressing is a hybrid targeting strategy: it is mostly triggered by humans, but also exploits platform infrastructure (notification systems, algorithmic curation). \n\nSources use the following strategies to target specific recipients or accounts directly:\n\n* #### @mentions\n* #### tagging (in photos, posts, or threads)\n* #### quote-tweets & reply-mentions\n* #### group direct messages\n* #### mailing-list addressing (To, CC)\n\nDirect addressing produces two distinct visibility effects:\n\n### Immediate Notification for the Targeted User {{notification_notifications}}\n\nThe targeted user receives the information item directly via notification, regardless of whether they would otherwise have encountered it through their feed.\n\n### Indirect Reach via Targeted Account’s Network\n\nWhen high-reach accounts are addressed, the information item can spread beyond the original notification: \n\n* visibility in the addressed account’s profile without any engagements:\\non some platforms (e.g. X/Twitter), @mentions are publicly visible in the addressed account's profile; \n* visibility to the entire account’s reach, i.e. their follower network:\\nwhen the addressed account replies, reposts, or reacts, the item reaches their entire follower network.\n\n## Content Adaptation to Platform Ranking Systems {{platform ranking system_platform ranking systems_content adaptation to platform ranking system_content adaptation}}\n\nSources adapt their content, metadata, and structure to fit the ranking criteria of the platforms on which they seek visibility — without paying for placement.\n\n### Search Engine Optimisation (SEO) {{search engine optimisation}}\n\nSEO refers to source-side practices that adapt websites — their content, metadata, and link structure — so that they rank more prominently in general-purpose search engine results (Google, Bing, DuckDuckGo, etc.).\n\nSEO is the most formalised optimisation strategy because search engine ranking signals are relatively stable and well-documented (Lewandowski et al.). Specialised SEO professionals, agencies, and tools support its practice.\n\n#### ☑ SEO Practices {{SEO practice_search engine optimisation practice_search engine optimisation practices}}\n\n\n:::success\n* keyword research and integration in titles, headings, body text\n\n\n* metadata optimisation (title tags, meta descriptions, alt text)\n* link building (acquiring inbound links from authoritative sites)\n* site structure and internal linking\n* page speed and mobile-friendliness\n* producing content that matches search intent\n\n:::\n\n\n:::info\nSEO does not change how search engines rank pages, it adapts the website to fit existing ranking criteria. Sources can shape what the algorithm sees, not how it decides.\n\n:::\n\n### Social Media Optimisation (SMO) {{social media optimisation_social media optimisations}}\n\nSMO refers to source-side practices to maximise visibility, engagement, and shareability of content on social media platforms.\n\nSMO is less formalised than SEO because social media ranking signals are more opaque and platform-specific. Practices shift as algorithms change.\n\n#### ☑ SMO Practices {{SMO practice_social media optimisation practices_social media optimisation practice}}\n\n\n:::success\n* hashtag strategies (trending or topic-specific tags)\n\n\n* posting timing (when target audiences are active)\n* content format choices (short video, carousels, reels)\n* headline and hook design (catching attention quickly)\n* encouraging engagement (questions, polls, calls to action)\n* cross-platform repurposing of content\n\n:::\n\n### Platform-Specific Optimisation {{platform-specific optimisations}}\n\nBeyond general SEO and SMO strategies, each platform follows its own ranking logic and content conventions, requiring tailored optimisation strategies.\n\n#### ☑ Platform-Specific Optimisation Strategies {{platform-specific optimisation strategy}}\n\n\n:::success\n* **TikTok**\n * Format: vertical short-form video, trending sounds\n * Strategies:\n * hook viewers in the first three seconds\n * post consistently during peak hours\n* **Instagram**\n * Format: high-quality visuals, Reels-first content\n * Strategies:\n * mix hashtags strategically\n * repurpose content across formats (posts, Stories, Reels)\n* **YouTube**\n * Format: long-form video with strong opening, keyword-rich titles and descriptions\n * Strategies:\n * optimise thumbnails for click-through\n * maximise watch time to trigger recommendations\n* **LinkedIn**\n * Format: long-form professional posts, native articles\n * Strategies:\n * publish natively (avoid external links)\n * encourage networked engagement through tagging\n* **X (Twitter)**\n * Format: concise text, threads for longer content\n * Strategies:\n * use strong hooks in the first line\n * reply to high-reach accounts to gain visibility\n\n:::\n\n\n:::info\nPlatform-specific optimisation requires understanding each platform's ranking system, audience behaviour, and content format preferences. What works on TikTok rarely works on LinkedIn.\n\n:::\n\n## Paid Visibility\n\n### Influencer Partnerships {{influencer partnership_influencer_influencers}}\n\nInfluencer partnerships are paid collaborations in which sources commission individuals with high account reach to promote their information items. \n\nPayment takes the form of monetary compensation or payment in kind (free products, services, or other benefits). \n\nThe core mechanism is reach-as-a-service: the source purchases access to the influencer's follower network. \n\n\n:::info\nInfluencer partnerships are sometimes clearly disclosed (\"Ad\", \"Paid partnership\"), sometimes only weakly or not at all. Disclosure standards vary by jurisdiction and platform.\n\n:::\n\n### Paid Placements {{paid placement_ads_ad_advertisements_sponsored content}}\n\nPaid placements are a source-side practice of paying platforms directly for visibility. \n\nSponsored content is placed alongside organic content in the form of \n\n* sponsored search results (search engine ads), \n* sponsored posts and promoted content (social media), or \n* display ads (banners, videos).\n\nPaid placements bypass organic ranking systems: instead of optimising content to rank well, the source pays the platform directly for placement.\n\n\n:::info\nThey are sometimes clearly labelled (\"Sponsored\", \"Ad\"), but are sometimes only weakly distinguishable from organic results. Labelling standards vary by jurisdiction and platform. \n\n:::\n\n# Platform-Side Curation\n\nPlatform-side curation is the platform’s own shaping of how prominently information items appear. It operates in two modes: \n\n* post-publication editorial curation \n* algorithmic curation\n\n## Post-Publication Editorial Curation\n\nEditorial curation, i.e. human editors or platform operators deliberately featuring selected items, plays only a minor role today: on most large digital platforms, prominence is determined overwhelmingly by algorithms. \n\nWhere it survives, it is largely confined to a few contexts, such as:\n\n* top-stories sections in news discovery systems (e.g. Apple News, Google News \"Editors’ picks\")\n* highlighted user posts in social media (e.g. LinkedIn’s in-house news team selecting \"Top Stories\" from member posts)\n\n## Algorithmic Curation\n\nAlgorithmic curation is the algorithmic selection, ranking, and presentation of available information items. It shapes how prominently they appear, not whether they are available at all (→ gatekeeping).\n\nIt operates in two modes, distinguished by whether the platform can recognise the individual user through a login, cookies, or device identifiers:\n\n* **non-personalised algorithmic curation:** the same for every user\\nIt draws only on information item and source properties and on aggregate engagement (the interactions an information item attracts across the whole audience), not on who is viewing.\n* **algorithmic personalisation:** tailored to the individual\\nIt draws on the user's tracked signals. If the platform can re-identify the user across visits and stores these signals, it builds an algorithmic profile of the user; the more this profile holds, the more closely the platform can tailor what the user sees.\n\n\n:::info\nWithout such recognition, for instance in a guest or incognito session without login or cookies, only non-personalised algorithmic curation applies. A platform may still adjust results by region through the IP address, but that is contextual, not individual: everyone in the same region sees the same thing.\n\n:::\n\n\n:::warning\n#### User reidentification\n\nThe platform recognises a returning user by one of three means, which differ in reach and persistence:\n\n* **login:** the user signs in with an account. This is the most reliable means and works across devices, because activity is tied to the account rather than to a single browser.\n* **cookies:** the platform stores an identifier in the user's browser and reads it on return. This is tied to one browser on one device and can be cleared by the user.\n* **device identifiers:** the platform reads technical features of the device or browser, or combines many of them into a near-unique fingerprint. This works even without a login or cookies and is the hardest for the user to escape.\n\nIn practice, platforms usually combine these means rather than relying on a single one.\n\n:::\n\n#### ☑ User Signals {{tracked signals_tracked user signals_tracked signal_tracked user signal_user signal}}\n\nUser signals are the data a platform tracks about a user to personalise the selection and visibility of information items. Many are generated as users select, interact with, respond to, or create items. Not all come from deliberate acts such as clicking or liking; some come from less deliberate behaviour, such as how long someone stays on a page or how far they scroll. Beyond these engagement-derived signals, platforms also draw on account and contextual data. From all of these they infer user interests, item relevance, user satisfaction, popularity, or likely future engagement. Users are often unaware how many signals may shape what they may encounter next.\n\n| Type | What it is | Examples |\n|------|------------|----------|\n| #### Engagement-derived signals | Signals captured during engagement, including attention measurements | \\- which items a user engaged with- the type of user action (→ engagement)
- measurements derived from the user action
(dwell time; scroll depth, hover / skip behaviour) |\n| #### Account data | Information about the user and their connections | \\- user profile (age, interests, profession, gender)
- language settings
- linked accounts
- contact list / address book |\n| #### Contextual data | Information about the situation in which the user is accessing the platform | \\- location, e.g. derived from the IP address- device type (e.g. phone or laptop)- time of access |\n\n\n:::info\nUnlike account and contextual data, engagement-derived signals depend on an information item's prior visibility: an item must first be shown before users can engage with it. This makes engagement self-reinforcing: an item with high prior engagement is ranked higher, gains more visibility, and so draws yet more engagement. This self-reinforcing dynamic is → algorithmic amplification.\n\nThe same dynamic can also be triggered deliberately by coordinated actors manufacturing engagement (→ coordinated amplification).\n\n:::\n\n\n:::info\n* Adomavicius, G., & Tuzhilin, A. (2011). Context-aware recommender systems. In F. Ricci, L. Rokach, B. Shapira, & P. B. Kantor (Eds.), *Recommender Systems Handbook* (pp. 217–253). Springer. \n* Kelly, D., & Teevan, J. (2003). Implicit feedback for inferring user preference: A bibliography. *ACM SIGIR Forum, 37*(2), 18–28. \n* Li, W., Kuo, J.-C., Sheng, M., Zhang, P., & Wu, Q. (2025). Beyond explicit and implicit: How users provide feedback to shape personalized recommendation content. In *Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25)*. Association for Computing Machinery. \n* Narayanan, A. (2023). *Understanding social media recommendation algorithms.* Knight First Amendment Institute, Columbia University. \n\n:::\n\n### Non-Personalised Algorithmic Curation\n\nNon-personalised algorithmic curation applies the same logic to every user, independent of who is searching or scrolling. It operates through two principal operations: \n\n* ranking\n* trending detection\n\n#### Ranking {{rank_order_rankings_ordering}}\n\nRanking arranges available information items into a ranked list, placing some higher and others lower on the visibility scale. The order is determined by ranking signals, typically combined rather than applied in isolation:\n\n* **information item and source properties**\n * **relevance:** how well an information item matches the query or context\n * **recency:** how current an information item is\n * **source authority:** how credible or high-quality a source is treated as\n* **aggregate audience response**\n * **engagement:** the interactions an information item attracts (likes, shares, comments, clicks) \n\n\n:::warning\nAlgorithms do not measure **source authority** directly; there is no single authority score. The signals are proxies from which it is estimated. Their type is documented through platform statements, patents, and industry observation, but their exact weighting is not independently verified, as platforms do not disclose their ranking systems.\n\n:::\n\n#### ☑ Signals Algorithms Use to Estimate Source Authority\n\n\n:::success\n#### **Link-based signals** \n\n* quantity and quality of inbound links from trusted sources\n* diversity of the backlink profile\n* links from topically related sources\n\n**Entity & brand reputation** \n\n* recognition as a named entity across the web\n* mentions in trusted sources (with or without a link)\n* volume of branded searches\n\n**E-E-A-T proxies** (Experience, Expertise, Authoritativeness, Trustworthiness)\n\n* identifiable authorship and stated credentials\n* publisher transparency (legal notice, ownership, contact)\n* secure site infrastructure (HTTPS)\n* factual accuracy and internal consistency\n\n**Domain-level track record**\n\n* domain age and continuity of focus\n* site-wide content quality\n* established reputation within the field\n\n**Citation-based signals** (academic databases, news) \n\n* citation count and quality of citing sources\n* journal rank or outlet standing\n\n:::\n\n#### Trending Detection {{trend detection_detecting trends}}\n\nTrending detection identifies topics or hashtags that are attracting a sudden surge of activity within a short period.\n\nIt operates on a different level from ranking. Ranking works on individual information items and arranges them into an ordered list. Trending detection does not order items: it reads the combined activity gathering around a topic or hashtag, across the many separate posts that address it or carry it, and highlights that topic or hashtag as a whole. It is a summarising view, surfaced in a dedicated section, such as:\n\n* trending topics and hashtags\n* a platform's \"what's happening\" or \"today's headlines\" section\n\nThe distinguishing signal is the rate of increase within a short window: how fast activity around the topic or hashtag is rising, not how much there is in total.\n\n### Algorithmic Personalisation\n\nAlgorithmic personalisation adapts the order, selection, or presentation of information items to individual users based on their tracked signals and algorithmic profiles:\n\n#### Personalised Ranking {{personalised rankings}}\n\nPersonalised ranking orders the items a user encounters differently for each individual, drawing on their tracked signals and algorithmic profiles:\n\n* personalised search results (shaped by location, history, profile)\n* social media feed ordering (\"For You\" feeds, \"Top posts\")\n\n#### Personalised Recommendations {{personalised recommendation_recommendation}}\n\nRecommendations suggest items beyond what the user actively requested:\n\n* \"Recommended for you\" video lists, \"Up next\" queues\n* suggested accounts, groups, or topics to follow\n* \"People you may know\"\n* related articles, similar products\n\n#### Personalised Advertising {{personalised advertisement}}\n\nPersonalised advertising targets paid content at individual users based on their algorithmic profile. Platform algorithms decide which advert reaches which user:\n\n* search ads tailored to past queries\n* social media sponsored posts based on profile and behaviour\n* retargeted display ads on websites\n\n\n:::warning\nTwo users on the same platform — even with the same query — typically see substantially different content.\n\n:::\n\n# Information Amplification {{information amplification_amplify_amplifies_amplification}}\n\nAmplification is the process by which the visibility of an information item, topic, or hashtag is broadened beyond the individual user to many accounts, and sometimes carried across information environments.\n\nThis produces three phenomena:\n\n* **trending**, where a topic or hashtag draws a surge of activity and the platform flags it with a dedicated slot;\n* **virality**, where a single item is passed on by so many accounts that it keeps reappearing;\n* **spill-over effect**, where the same information comes to appear in different items across more than one information environment.\n\nWithin an information environment, two mechanisms drive amplification:\n\n* **Algorithmic amplification** is platform-driven: \\nthe engagement feedback within → platform-side curation, which keeps surfacing an already-visible item to more accounts.\n* **Coordinated amplification** is actor-driven: \\nmultiple accounts, groups, or campaigns act in concert to boost visibility beyond what individual activity would produce. The literature distinguishes it on two axes, whether the coordination is transparent or concealed and whether the accounts are real or fake (Rogers & Righetti, 2025):\n * **Coordinated authentic amplification**: \\ntransparent coordination, real accounts (e.g. open civic campaigns, advocacy, marketing).\n * **Coordinated inauthentic or artificial amplification**: \\nconcealed coordination, fake accounts, or both, manufacturing an appearance of organic support (Meta's Coordinated Inauthentic Behaviour / CIB; Gleicher, 2018).\n\n\n:::info\nThe two mechanisms frequently combine: coordinated networks exploit engagement-based ranking to trigger algorithmic boosts, and algorithmic ranking in turn compounds whatever visibility the coordination has produced. \n\nSpill-over works differently: through → re-creation, an actor carries information across an environment boundary, by journalistic reporting, cross-platform sharing, or editorial pickup, rather than amplifying it within one.\n\n:::\n\n\n:::info\n* Gleicher, N. (2018). *Coordinated Inauthentic Behavior Explained*. Meta Newsroom. \n\n\n* Rogers, R., & Righetti, N. (2025). Coordinated inauthentic behaviour on Facebook? A typology of manufactured attention. \n\n:::\n\n## Algorithmic Amplification {{algorithmic amplification}}\n\nAlgorithmic amplification is a self-reinforcing dynamic that amplifies algorithmic curation: because engagement-derived signals depend on an item's prior visibility, an item with high prior engagement is ranked higher, gains more visibility, and so draws yet more engagement (the rich-get-richer effect). Because engagement is the trigger, the dynamic can also be set off deliberately through manufactured engagement (→ coordinated amplification).\n\nEmpirical research shows that engagement-based ranking systematically amplifies emotionally charged and out-group hostile content, even when users themselves do not prefer such content (Milli et al., 2025). \n\nAlgorithmic amplification is not a neutral reflection of user activity. Its effects are emergent and visible primarily in the aggregate: individual recommendations are imprecise (engagement rates remain below 1% on most platforms), but ranking, recommendation, and demotion systematically shape what circulates across the platform.\n\n\n:::info\n* Milli, S., et al. (2025). Engagement, user satisfaction, and the amplification of divisive content on social media. PNAS Nexus.\n\n\n* Narayanan, A. (2023). Understanding social media recommendation algorithms. Knight First Amendment Institute.\n\n:::\n\n## Coordinated Authentic Amplification {{coordinated authentic amplifications}}\n\nCoordinated Authentic Amplification is the deliberate boosting of an information item, topic, hashtag, account, or viewpoint through openly disclosed, organised activity by real accounts. The coordinated origin is not concealed: participants act under their real identities or under known group affiliations.\n\nTypical contexts include \n\n* civic campaigns (e.g. NGO petitions, advocacy hashtags), \n* political mobilisation (e.g. party campaigning, get-out-the-vote efforts), \n* marketing and brand campaigns, \n* professional association communications, and \n* cultural movements such as Fridays for Future or #MeToo.\n\n\n:::warning\nWhether the underlying message is well-founded, balanced, or one-sided is a separate question — authenticity refers only to the transparency of the coordination, not to the truth-value or fairness of the content. An authentic campaign can amplify accurate information, misleading information, or a one-sided position.\n\n:::\n\n\n:::warning\nAuthentic and inauthentic coordination can produce visibility patterns that look identical from the outside — synchronised sharing, hashtag clustering, rapid uptake. The distinguishing feature is not the visible pattern but whether the coordinated origin is openly disclosed.\n\n:::\n\n## Coordinated Inauthentic or Artificial Amplification {{coordinated inauthentic amplification_coordinated inauthentic amplifications_coordinated artificial amplification_coordinated artificial amplifications}}\n\nCoordinated inauthentic or artificial amplification is the deliberate boosting of an information item, topic, hashtag, account, or viewpoint through organised activity in which the coordinated origin is concealed, the participating accounts are fake, or both. The aim is to manufacture an appearance of organic, independent support. Meta's term *Coordinated Inauthentic Behaviour* (CIB) — now incorporated into the EU Digital Services Act — centres on this combination of false identities and adversarial methods to evade detection (Gleicher, 2018; Rogers & Righetti, 2025).\n\nTypical contexts include \n\n* political influence operations (state-sponsored or party-aligned), \n* astroturfing campaigns (commercial or ideological), \n* targeted disinformation around elections, public health, or geopolitical conflict, and \n* reputation manipulation through fake reviews, ratings, or engagement. \n\nThe operational means — bots, trolls, sockpuppets, and their coordinated networks (bot farms, troll farms, sockpuppet networks, click farms) — are described in detail below.\n\n\n:::warning\nInauthenticity refers to the concealment of the coordinated origin or the use of fake accounts — not to the truth-value of the content being amplified. A coordinated network of fake accounts can amplify accurate information; a single authentic individual can spread fabricated information. Coordinated inauthentic amplification and the spread of false content are distinct phenomena that can occur independently or together.\n\n:::\n\nThe following account types described in this section apply across digital information channels & platforms where users can create accounts and post or interact publicly — particularly social media, discussion forums and community spaces, video and audio platforms, and review or comment sections. They are less prominent in private communication apps or in environments without user-generated content.They appear both independently and within coordinated networks. They are listed here because of their typical role in amplification dynamics; the explicitly coordinated formations are the Account Networks.\n\n| Term | Definition | Controlled by | Defined by | Typical purpose |\n|------|------------|---------------|------------|-----------------|\n| **Social Bot** | An automated or partly automated account that posts, likes, follows, shares, or replies online. | Software | **Automation** | To amplify messages, create artificial popularity, spam, influence debate, or spread content at scale. |\n| **Cyborg** | A hybrid account combining human operation with software automation. | Mixed: human and software | **Selective automation** | To combine the scale of automation with the contextual plausibility of human input — for legitimate scheduling/management or for harder-to-detect influence operations. |\n| **Troll** | A person or account that deliberately provokes, disrupts, or inflames online discussion. | Usually a human user; sometimes coordinated groups | **Disruptive / provocative / antagonistic behaviour** | To upset others, derail conversations, provoke reactions, spread hostility, or polarise debate. |\n| **Sockpuppet** | A fake account used by someone to hide their real identity or pretend to be a different person. | A human user, though the account may also use automation | **Deceptive identity** | To create false support, attack others anonymously, evade bans, manipulate debate, or give the impression of independent agreement. |\n\n### Social Bot {{social bots}}\n\nA **social bot** is a bot designed to operate on social media platforms, posting, commenting, sharing, or interacting in ways that simulate human users. Social bots are typically programmed to act at scale and at high speed, far beyond what a human user could manage. Their activity is often repetitive and coordinated across many accounts, which distinguishes it from normal human use.\n\nSocial bots can be used for legitimate purposes — such as customer service, news distribution, or marketing — but they are also widely used to influence public opinion, amplify certain messages, manipulate discussions, or manufacture the appearance of widespread support for specific ideas, products, or causes. In the context of misinformation and disinformation, social bots play a particular role in spreading content rapidly and giving the false impression that many independent voices share the same view.\n\nWhen social bots are deployed in coordinated networks, they form a *Bot Farm*.\n\n### Cyborg {{cyborgs}}\n\nA **cyborg** is a hybrid account that is partly operated by a human and partly automated by software. A cyborg may have routine posts scheduled or generated by software while a person handles selected interactions, replies, or sensitive content. The balance between automated and human activity varies between accounts.\n\nCyborgs can be used for legitimate purposes — such as content scheduling, brand or institutional account management, or hybrid customer service — but they are also used in influence operations to combine the scale and speed of automation with the contextual plausibility of human input.\n\nCyborgs are more difficult to identify than purely automated bots because part of their behaviour is genuinely human, which means single detection indicators rarely suffice for reliable identification.\n\n### Troll {{trolls}}\n\nA **troll** is a real person who deliberately disrupts online discussions through provocative, aggressive, or hostile behaviour. Trolls typically use personal accounts and target controversial issues, public figures (such as politicians or journalists), or media organisations. Their aim is to upset others, trigger reactions, or escalate conflicts — sometimes in support of a particular agenda, sometimes for entertainment or attention.\n\nWhile trolls often act independently, they may also operate in coordinated groups, sometimes paid by political or commercial actors (see *Troll Farm* under Mechanisms of Amplification).\n\n**Trolling is best understood as a pattern of online behaviour, not a specific kind of account.** The same behaviour can be performed by automated accounts, and ordinary users can engage in trolling on occasion.\n\n#### Sockpuppet {{sockpuppets}}\n\nA **sockpuppet** is a fake online identity created and operated by a real person who hides their true identity. Unlike trolls — who often act under a single openly hostile account — a sockpuppet operator typically runs multiple fake accounts in parallel to create the impression that several independent users hold the same opinion, support the same cause, or agree with the operator's own (often separate) main account.\n\nSockpuppets are commonly used to manufacture artificial consensus, support one's own arguments under different names, attack opponents while appearing impartial, evade bans by creating new identities after suspension, or manipulate online reviews, votes, and polls.\n\nSockpuppets differ from social bots in that they are manually operated by humans, which makes their content more contextually plausible and harder to detect through automated means. They differ from trolls in that their primary goal is deception about identity and the manufacturing of apparent consensus, not provocation — although sockpuppet operators can also engage in trolling behaviour through their fake identities.\n\nWhen a person or small group operates a coordinated set of sockpuppets together, they form a *Sockpuppet Network* (see Mechanisms of Amplification).\n\n#### ☑ Differentiating Social Bots, Trolls, and Sockpuppets {{differentiating social bots_differentiating trolls_differentiating sockpuppets}}\n\n| **Detection Dimension** | **Social Bots** | **Trolls** | **Sockpuppets** |\n|---------------------|-------------|--------|-------------|\n| **Profile Characteristics** | - [ ] The account looks newly created
- [ ] The profile is incomplete or generic
- [ ] The username may look non-personal and sometimes include random numbers | - [ ] The account has typically been active for longer and has a post history
- [ ] The profile is complete and seems personal; it may present strong ideological or political self-description
- [ ] The username looks personal | - [ ] The profile looks plausible and personal, often with a profile picture and biographical details (sometimes stolen, AI-generated, or copied)
- [ ] Account history may be moderate and designed to look authentic over time |\n| **Posting Behaviour** | - [ ] The activity does not match normal human online behaviour
- [ ] The accounts post or repost content very frequently
- [ ] The accounts post or repost content at all hours, day and night | - [ ] The activity resembles normal human online behaviour
- [ ] The account posts or replies at irregular times
- [ ] The account becomes more active during controversial discussions | - [ ] Activity patterns resemble normal human use
- [ ] Multiple accounts run by the same operator may show similar active hours or rhythms
- [ ] Sockpuppets tend to start fewer discussions and write shorter posts than typical users |\n| **Interactions** | - [ ] The account does not have real conversations
- [ ] The accounts mostly like, share, or repost
- [ ] Replies are short and automated | - [ ] The account replies directly to other users
- [ ] The account engages in debates with the purpose of provoking reactions
- [ ] Conversations are extended to create or escalate conflict | - [ ] The account engages in real conversations, often supporting the operator's main account or other sockpuppets
- [ ] Replies are contextually appropriate and seem authentic
- [ ] Pairs of sockpuppets often interact in the same discussion at similar times |\n| **Content Features** | - [ ] The content is one-sided and repetitive
- [ ] The same narratives are posted many times | - [ ] The content is specifically tailored to harm or provoke a target
- [ ] The content targets individuals or social groups | - [ ] Content seems genuine and varied across accounts
- [ ] The underlying message or stance aligns suspiciously across the network
- [ ] More frequent use of personal pronouns such as \"I\" |\n| **Language** | - [ ] Generic expressions, repetitive phrasing with keywords | - [ ] Varied, emotional, often abusive or offensive language | - [ ] Natural and varied language
- [ ] Multiple accounts may share linguistic fingerprints (similar phrasing, vocabulary, punctuation, or error patterns) |\n| **Network & Technical Indicators** | - [ ] Social bots follow other social bots, but the relationship is typically one-way and not reciprocal
- [ ] Coordinated behaviour is observable across multiple bot accounts | - [ ] Trolls follow human accounts
- [ ] Connections are often reciprocal (they follow their followers and vice versa)
- [ ] Trolls typically act independently of each other | - [ ] Multiple accounts engaging with each other in mutually supportive ways
- [ ] Connections may be artificially reciprocal between sockpuppets in the same network, or deliberately absent to avoid detection
- [ ] Same IP address, device fingerprint, or login pattern \\\\\\*(platform-side detection)\\\\\\*
- [ ] More clustered ego-networks than ordinary users
- [ ] Correlated activity timing across accounts |\n\n\n:::info\n* Ferrara, E. (2023). Social bot detection in the age of ChatGPT: Challenges and opportunities. *First Monday, 28*(6). \n* Kumar, S., Cheng, J., Leskovec, J., & Subrahmanian, V. S. (2017). An army of me: Sockpuppets in online discussion communities. *Proceedings of the 26th International Conference on World Wide Web (WWW '17)*, 857–866. \n* Orabi, M., Mouheb, D., Al Aghbari, Z., & Kamel, I. (2020). Detection of bots in social media: A systematic review. *Information Processing & Management, 57*(4), 102250. \n* Solorio, T., Hasan, R., & Mizan, M. (2013). A case study of sockpuppet detection in Wikipedia. *Proceedings of the Workshop on Language Analysis in Social Media (LASM) at NAACL-HLT*, 59–68. Association for Computational Linguistics. \n* Tomaiuolo, M., Lombardo, G., Mordonini, M., Cagnoni, S., & Poggi, A. (2020). A survey on troll detection. *Future Internet, 12*(2), \n* Tsikerdekis, M., & Zeadally, S. (2014). Multiple account identity deception detection in social media using nonverbal behavior. *IEEE Transactions on Information Forensics and Security, 9*(8), 1311–1321. \n* Uyheng, J., Moffitt, J. D., & Carley, K. M. (2022). The language and targets of online trolling: A psycholinguistic approach for social cybersecurity. *Information Processing & Management, 59*(5), 103012. \n\n:::\n\n### Account Networks {{account network}}\n\n#### Bot Farm {{bot farms}}\n\nA bot farm is a network of bots operating simultaneously across multiple devices or servers, deployed by a single operator or organisation for a particular purpose.\n\nBot farms have a range of legitimate uses, including web indexing, automated software testing, data aggregation, and website performance monitoring. However, they are also commonly used for malicious activities such as creating fake engagement, generating large volumes of content, distributing spam, or carrying out cybersecurity attacks. When used to manipulate online discourse, bot farms can create the false impression of widespread support, opposition, or interest in a topic, account, or campaign.\n\n#### Troll Farm {{troll farms}}\n\nA troll farm is an organised group of coordinated, often paid workers who post deliberately provocative, misleading, or false content online — typically through fake accounts. Their aim is usually to manipulate public opinion, spread disinformation, or create social and political unrest. Troll farms have been documented in connection with state-sponsored influence operations as well as commercial reputation manipulation.\n\n#### Sockpuppet Network {{sockpuppet networks}}\n\nA sockpuppet network is a coordinated set of sockpuppet accounts operated by one person or a small group, used to simulate independent voices supporting a shared narrative, campaign, account, or cause. Sockpuppet networks are commonly used in political astroturfing, review and rating manipulation, and coordinated disinformation campaigns. Unlike bot farms, sockpuppet networks rely on manual operation by humans, which makes the content of individual accounts appear more authentic and harder to detect through automated means. Their coordination usually becomes detectable only when multiple accounts can be linked through behavioural patterns, shared technical signals, or mutual engagement.\n\n#### Click Farm {{click farms}}\n\nA click farm is an operation where large numbers of low-paid workers, automated bots, or both are used to click on ads, follow social media accounts, like posts, leave reviews, or download apps. The goal is to artificially boost online engagement or traffic, making content, accounts, or products appear more popular than they actually are.\n\n# Information Narrowing\n\nNarrowing is the process by which the range of information items and viewpoints reaching an individual user or group is restricted to what already fits their profile or shared beliefs, while diverging content gradually drops away. Two mechanisms drive narrowing, each producing one phenomenon:\n\n* **Algorithmic narrowing** is platform-driven: \\na self-reinforcing feedback loop of algorithmic personalisation keeps tailoring what a user sees to their tracked signals. Over repeated cycles the user may become enclosed in a **filter bubble**, where content that diverges from their profile seldom gets through.\n* **Social reinforcement** is actor-driven: \\nmembers of a self-chosen group affirm viewpoints that fit their shared beliefs and dismiss those that diverge, so the group's range of viewpoints contracts. The group settles into an **echo chamber**, where members mostly encounter their own position reflected back.\n\n## Algorithmic Narrowing\n\n#### Algorithmic Narrowing\n\nAlgorithmic narrowing is the platform-side mechanism behind the filter bubble. A self-reinforcing feedback loop of algorithmic personalisation tailors what a user sees to their tracked signals: the user engages mostly with what already fits their profile, that engagement sharpens the profile, and each new selection is tighter than the last. Through these repeated cycles, the range of items a user meets steadily contracts.\n\nThe filtering is not done by the user or a chosen group, as in social reinforcement, but by the platform itself: the user need not seek out like-minded content; the personalisation system narrows the selection on its own, often unnoticed.\n\nResearch on recommender systems shows that such feedback loops degenerate over time: the diversity of items a user engages with narrows, and their interests drift towards the extreme (Jiang et al., 2019). Unlike algorithmic amplification, which runs on an item's prior visibility and can push content a user does not prefer, this loop runs on the user's own engagement and therefore follows their preferences rather than overriding them.\n\nThe resulting state and its consequences are covered under filter bubble.\n\n## Social Reinforcement\n\nSocial reinforcement is the social mechanism behind the echo chamber. Within a self-chosen group, members affirm and reward viewpoints that fit shared beliefs, while viewpoints that diverge are dismissed or excluded. Through this mutual response, the range of perspectives a member meets steadily contracts.\n\nThe filtering is not done by the platform, as in algorithmic personalisation, but by the group itself: no automated system removes diverging views; the members do, through how they respond to one another.\n\nThe resulting state and its consequences, including group polarisation and the distinction from an epistemic bubble, are covered under echo chamber.","HTML":"

Uebersetzung konnte aktuell nicht geladen werden.

","UPDATEDAT":"2026-07-10T09:41:41.611Z","LANG":"sl","ID":"b9d42740-38bc-4e99-8a15-b37d5c54a10d","DEEPLRETRYAFTER":"2026-08-19 16:12:25","TITLE":"Information Visibility in Digital Information Environments","SOURCELANG":"en"}