Branchoria Think Before Sharing Think Before Sharing

Practical judgement for social feeds, AI answers, and viral claims.

Within Think Before Sharing

Why Did This Claim Find You?

Critical thinking online starts with asking why a post reached you, not only whether the post itself is true.

On this page

  • Feeds versus searches
  • Signals that shape ranking
  • Questions to ask about distribution
Preview for Why Did This Claim Find You?

Introduction

A claim online is not only something to verify; it is also something that was delivered. Critical thinking in social media and AI environments starts by asking, “Why did this claim find me?” A post may appear because you searched for it, but more often it arrives through ranking systems that predict what will keep you watching, clicking, commenting, sharing or returning. That means the path to your screen is part of the evidence.

Overview image for Ranking Algorithmic ranking does not make content automatically false, manipulative or dangerous. It can help people discover useful reporting, niche expertise, art, education and community. But it changes the first critical question. Instead of treating a feed as a neutral window onto “what people are saying”, it is more accurate to see it as an ordered selection produced by signals, predictions, business incentives, safety rules and design choices. You are not just judging a claim; you are judging a claim that has already survived a distribution contest.

Feeds Are Not Searches

A search begins with your question. A feed often begins with a platform’s guess. That difference matters because search and feed ranking create different habits of attention. In a search, the user usually knows they are seeking information and can compare results against an intention: “Which result answers my query?” In a social feed, the question is often supplied after the fact: a clip, screenshot or claim appears, and only then does the user decide whether it matters.

YouTube’s own explanation makes this distinction visible. It says recommendations appear across the homepage, Up Next, Shorts and other surfaces, and that the homepage is primarily personalised. It also says YouTube learns from watch history, search history, subscriptions, likes, dislikes, “Not interested” feedback and “Don’t recommend channel” feedback. In other words, even when a user is not searching, past behaviour can shape what appears next. [Google Help]support.google.comHelp How You Tube recommendations workGoogle HelpHow YouTube recommendations work - YouTube Help…

TikTok describes the same general mechanism in its own terms: a recommender system selects from eligible content and ranks it according to predictions about what a user is likely to be interested in. It may also use patterns from people with similar interests, and it says feeds such as For You, Following, Friends and LIVE can remain unique to each person. [TikTok Support]support.tiktok.comTik Tok Supportsupport.tiktok.comTik Tok Supportsupport.tiktok.com

That is why the phrase “the algorithm showed me” is not a complete explanation. A feed may include material because it matches your behaviour, because similar users reacted to it, because it is rising quickly, because it fits a platform’s policy rules, because it fills a diversity slot, because a creator paid to promote it, or because the system is testing whether you will engage. The post is the visible object. The ranking process is the hidden route.

What Ranking Systems Are Trying to Predict

Most large social platforms do not simply sort posts by time. They estimate the likely value of showing one item rather than another. In practice, that “value” is usually a bundle of predicted actions: whether you will watch, pause, click, like, share, comment, follow, hide, report, subscribe or keep scrolling.

Meta says its AI systems predict how valuable a piece of content might be to a user so that it can be shown sooner, and gives sharing as one example of a signal that may indicate interest. It also says no single prediction is a perfect gauge, so ranking uses many predictions in combination, including behavioural signals and user feedback from surveys. [About Facebook]about.fb.comAbout Facebook How AI Influences What You See on Facebook and InstagramAbout Facebook How AI Influences What You See on Facebook and Instagram

This is the mechanism behind a crucial critical-thinking point: a highly ranked post is not necessarily the most accurate, important or representative post. It is a post that performed well against ranking criteria. Those criteria may overlap with quality, but they are not the same thing. A careful investigation, a misleading rumour, a funny clip and an angry reaction can all be “successful” in a ranking system if they produce strong signals.

Recommendation systems are therefore distribution engines, not truth engines. Arvind Narayanan’s explanation for the Knight First Amendment Institute puts the point plainly: when people post online, who hears them is determined in large part by recommender algorithms, and these systems are central to platforms such as Facebook, YouTube and TikTok. [Knight First Amendment Institute]knightcolumbia.orgOpen source on knightcolumbia.org.

Ranking illustration 1

Signals That Shape What Finds You

The exact ranking formula varies by platform and changes over time, but the broad signal categories are understandable. A reader does not need to know the source code to ask better questions about distribution.

Common ranking signals include: [brainforge.ai]brainforge.aiSource details in endnotes.

  • Your own behaviour: what you watch, search, like, share, pause on, replay, hide or report.
  • People like you: what users with similar patterns have watched or engaged with.
  • The content itself: topic, caption, hashtags, audio, language, format, account history, freshness and eligibility.
  • Social proof: how quickly others are reacting, whether the item is being shared, and what type of accounts are spreading it.
  • Relationship signals: whether you follow the creator, message them, visit their profile or interact with mutual connections.
  • Platform constraints: safety rules, quality demotions, diversity injections, local law, advertising systems and product priorities.

X’s public recommendation repository gives a rare glimpse of how many components can sit behind a feed. Its README describes the recommendation algorithm as services and jobs responsible for feeds across surfaces such as For You, Search, Explore and Notifications. It lists components for real-time user actions, explicit signals such as likes and replies, implicit signals such as profile visits and post clicks, graph-based relationship features, reputation calculations and trust-and-safety models. [GitHub]github.comGitHub - twitter/the-algorithm: Source code for the X Recommendation Algorithm · GitHub…

That complexity matters because users often interpret ranking as endorsement: “This is everywhere, so it must be important.” But a post can be everywhere in one person’s feed because of a narrow behavioural loop. If you watch three videos about a rumour, the platform may infer interest even if your actual motive was scepticism, anxiety or anger. The system observes behaviour; it cannot always infer why the behaviour occurred.

YouTube Recommendation Engine: Complete Meltdown Analysis

Channel: ByteMonk · Views: 28.5K · Uploaded: February 2026 · Length: 12 minutes

Open on YouTube

Why Engagement Can Misread Interest

The most important ranking mistake for critical thinking is confusing engagement with endorsement. A user may linger on a post because it is useful, but also because it is shocking, irritating, confusing, embarrassing, funny, threatening or hard to parse. A comment may be agreement, correction, mockery or outrage. A share may be recommendation, warning or ridicule.

This is why “I only watched because it annoyed me” still matters. Many systems learn from attention, not from your private judgement. You may think you are rejecting a claim, while your behaviour helps mark the topic, format or creator as relevant to you. Platforms do provide corrective controls, but those controls are not always obvious, and their effects can vary by surface.

Research on YouTube recommendation controls illustrates the point. A 2023 audit using simulated users found that watching a topic could increase its presence on the homepage, and that the “Not interested” button was the most effective tested method for reducing unwanted homepage recommendations, removing 88% on average in the topics tested. But the same study found that 44% of surveyed adult YouTube users in the US were unaware the button existed, and that the controls had much less effect on video-page recommendations. [arXiv]arxiv.orgarXiv How to Train Your You Tube Recommender to Avoid Unwanted VideosHow to Train Your YouTube Recommender to Avoid Unwanted VideosJuly 27, 2023…Published: July 27, 2023

Recent TikTok research points to a similar problem of user agency. A 2026 study found that TikTok’s For You Page was sensitive to both explicit and implicit signals, but that users could struggle to stop unwanted topics from returning; the most effective explicit signal, marking a video as “Not Interested”, was described as unintuitively buried in the interface. [arXiv]arxiv.orgWhen 'For You' Isn't For You: Measuring User Agency in TikTok's Algorithmic FeedMay 11, 2026…Published: May 11, 2026

The critical-thinking lesson is practical: passive irritation is still data. If a topic is unpleasant, misleading or compulsive, scrolling past may be a weaker signal than explicitly hiding it, reporting it where appropriate, clearing history, switching to a following-only feed, or leaving the platform surface entirely.

Distribution Is Evidence, But Not Proof

How a claim reached you can tell you something about its likely function. A claim arriving through a friend’s direct message, a search result, a paid advert, a creator’s series, a trending feed, a recommended short video and a screenshot in a group chat should not be read in the same way. The route does not prove truth or falsehood, but it gives clues about incentives and context.

A useful test is to separate three questions:

  1. What is the claim saying? Identify the concrete statement, not just the emotion around it.
  2. What evidence supports it? Look for original documents, named sources, dates, data, methods and independent confirmation.
  3. Why is this version being distributed to me now? Ask whether the ranking path rewards urgency, outrage, identity, spectacle, fear, humour or usefulness.

This matters especially when AI-generated content is involved. A synthetic image, summary or voice clip can be produced cheaply, but its real impact depends on distribution. A false post with no reach is a small problem; a false post tuned to travel through recommendation systems can become a public issue. The spread mechanism is part of the harm.

Ranking also affects perceived consensus. If a feed repeatedly shows the same claim, many users experience it as social evidence: “Everyone is talking about this.” But repetition in a personalised feed is not the same as broad agreement. It may reflect your inferred interests, a local cluster, coordinated posting, a trending format, or a feedback loop in which engagement creates more exposure and more exposure creates more engagement.

Ranking illustration 2

The Feed Can Narrow Without Feeling Narrow

Personalisation can feel expansive because it keeps producing new posts. Yet the variety may be shallower than it appears: different creators, sounds and formats can still point towards the same emotional or ideological groove. The user feels busy, informed or immersed, while the range of sources and interpretations quietly shrinks.

TikTok says it tries to diversify recommendations by introducing users to new creators and content, and says it may show material that does not appear relevant to expressed interests. It also says it generally avoids recommending already seen content and may limit or avoid recommending certain permitted content that is not suitable for a general audience. [TikTok Support]support.tiktok.comTik Tok Supportsupport.tiktok.comTik Tok Supportsupport.tiktok.com

But independent audits suggest the balance between reinforcement and exploration is hard to get right. A 2025 sock-puppet audit of TikTok reported strong amplification of content aligned with bots’ interests, with rapid reinforcement typically occurring within the first 200 videos watched; it also found a negative relationship between amplification and exploration, meaning that as interest-aligned content increased, engagement with unseen hashtags declined. [arXiv]arxiv.orgarXiv Dynamics of Algorithmic Content Amplification on Tik TokDynamics of Algorithmic Content Amplification on TikTokMarch 26, 2025…Published: March 26, 2025

For a reader, the takeaway is not “personalisation is bad”. It is that a personalised feed is a shaped environment. It may be excellent at finding more of what you have already signalled, but weaker at distinguishing curiosity from belief, concern from endorsement, or a temporary interest from a healthy long-term information diet.

Ranking Changes Journalism and Public Knowledge

Algorithmic ranking also changes the public information environment around the user. News is no longer encountered only through a newspaper homepage, evening bulletin or direct visit to a trusted outlet. It increasingly appears inside mixed feeds where journalism, influencer commentary, entertainment, advertising, activism, satire, scams and AI-generated material can share the same visual grammar.

The Reuters Institute’s 2025 Digital News Report describes traditional news media as struggling with declining engagement, low trust and stagnant digital subscriptions, while its executive summary highlights an accelerating shift towards social media and video platforms that is fragmenting the alternative media environment. [reutersinstitute.politics.ox.ac.uk]reutersinstitute.politics.ox.ac.ukdigital news reportdigital news report

This does not mean institutional journalism is always right or that creators are always unreliable. It means the old cues are weaker. A video can look professional without editorial checks. A local rumour can appear beside a legitimate council update. A screenshot can detach a quote from its date and source. A creator may be knowledgeable in one area and careless in another.

For critical thinking, the ranking question becomes: “Am I seeing the best available account, or the most feed-compatible account?” Feed-compatible information is often short, emotional, visual, personalised and easy to react to. High-quality information may be slower, less dramatic, less personalised and less likely to be rewarded by immediate engagement.

An inside look at X’s Community Notes | Keith Coleman & Jay Baxter

Channel: Lenny's Podcast · Views: 9.1K · Uploaded: February 2025 · Length: 1 hour 47 minutes

Open on YouTube

Platform Controls Help, But They Are Not Full Transparency

Platforms increasingly offer explanations and controls: “Why am I seeing this?”, “Not interested”, “Show less”, “Following”, “Favourites”, watch-history deletion, ad preferences and chronological feeds. These tools are worth using, but they are not the same as full user control.

Meta says it has expanded “Why Am I Seeing This?” explanations and centralised controls such as Feed Preferences, Suggested Content Control Centre, “Show more”, “Show less”, chronological feeds and favourites. It also says it releases system cards explaining how ranking systems work and what controls users can use. [About Facebook]about.fb.comAbout Facebook How AI Influences What You See on Facebook and InstagramAbout Facebook How AI Influences What You See on Facebook and Instagram

The European Union’s Digital Services Act has made this issue a legal requirement for many platforms. Article 27 requires online platforms using recommender systems to explain, in plain and intelligible language, the main parameters used and any options users have to modify or influence them. It also says platforms should explain why certain information is suggested and provide accessible functionality where users can select and modify available ranking options. [EU Digital Services Act]eu-digital-services-act.comEU Digital Services Act Article 27, the Digital Services Act (DSAEU Digital Services Act Article 27, the Digital Services Act (DSA

This is progress, but it has limits. A user-facing explanation may say that a post appeared because of prior activity, similar users, location or topic interest, without revealing the relative weight of each factor in the moment. Platforms also have legitimate reasons not to disclose every anti-abuse signal. The practical reader response is to treat controls as steering tools, not as a complete map of the machine.

How Algorithms Shape What You See Online

Channel: ExplainThatFast · Views: 27 · Uploaded: August 2025 · Length: 2 minutes 5 seconds

Open on YouTube

Questions to Ask When a Claim Finds You

A good ranking-aware reading habit is short enough to use in the moment. Before liking, sharing or repeating a claim, ask:

  • Did I ask for this, or was it selected for me? Search results, subscriptions and recommendations deserve different levels of caution.
  • What behaviour might have trained this feed? Recent watches, searches, pauses, comments and shares may explain why the claim appeared.
  • Is the post asking for evidence or asking for reaction? Urgency, anger and identity cues often travel well even when evidence is weak.
  • What is missing from the route? Look for the original source, date, location, full clip, full document or named expert.
  • Would I believe this if it appeared in a less flattering feed? Claims that confirm identity or prior belief need extra friction.
  • Is repetition making it feel verified? Seeing the same claim several times may show algorithmic reinforcement, not independent confirmation.
  • Can I change the distribution signal? Hide, mark “Not interested”, unfollow, report, clear history or switch feed mode when the system is learning the wrong lesson.

The aim is not to become suspicious of every post. It is to restore the missing context. A feed hides much of the selection process, so the reader has to put distribution back into the act of judgement.

Ranking illustration 3

The Core Habit: Read the Route as Well as the Claim

Algorithmic ranking turns attention into infrastructure. It decides which claims get repeated, which sources become familiar, which moods dominate a session, and which topics seem more socially important than they may be. That makes ranking a central part of critical thinking online.

The practical rule is simple: when a claim finds you, inspect the route. A claim that arrives through a recommendation system has already been filtered by predictions about engagement, relevance, similarity, safety and platform goals. Those predictions may help you discover something valuable, but they are not evidence of accuracy.

The strongest online readers therefore do two things at once. They check whether the claim is true, and they ask why this particular version of the claim reached them in this particular format at this particular moment. That second question is not a distraction from truth. In the age of social media and AI, it is one of the fastest ways back to it.

Amazon book picks

Further Reading

Books and field guides related to Why Did This Claim Find You?. Use these as the next step if you want deeper reading beyond the article.

Browse more on Amazon: The Chaos Machine Weapons of Math Destruction The Filter Bubble

As an Amazon Associate I earn from qualifying purchases.

eBay marketplace picks

Marketplace Samples

Example marketplace items related to this page. Use the search link to explore similar finds on eBay.

Using USA
Browse more on eBay.co.uk

Example items shown for inspiration; availability and pricing can change. Branchoria may earn a commission if you purchase through outbound eBay links.

Endnotes

  1. Source: support.google.com
    Title: Help How You Tube recommendations work
    Link: https://support.google.com/youtube/answer/16089387?hl=en
    Source snippet

    Google HelpHow YouTube recommendations work - YouTube Help...

  2. Source: support.tiktok.com
    Title: Tik Tok Supportsupport.tiktok.com
    Link: https://support.tiktok.com/en/using-tiktok/exploring-videos/how-tiktok-recommends-content

  3. Source: github.com
    Link: https://github.com/twitter/the-algorithm
    Source snippet

    GitHub - twitter/the-algorithm: Source code for the X Recommendation Algorithm · GitHub...

  4. Source: arxiv.org
    Title: arXiv How to Train Your You Tube Recommender to Avoid Unwanted Videos
    Link: https://arxiv.org/abs/2307.14551
    Source snippet

    How to Train Your YouTube Recommender to Avoid Unwanted VideosJuly 27, 2023...

    Published: July 27, 2023

  5. Source: arxiv.org
    Link: https://arxiv.org/abs/2605.10690
    Source snippet

    When 'For You' Isn't For You: Measuring User Agency in TikTok's Algorithmic FeedMay 11, 2026...

    Published: May 11, 2026

  6. Source: arxiv.org
    Title: arXiv Dynamics of Algorithmic Content Amplification on Tik Tok
    Link: https://arxiv.org/abs/2503.20231
    Source snippet

    Dynamics of Algorithmic Content Amplification on TikTokMarch 26, 2025...

    Published: March 26, 2025

  7. Source: reutersinstitute.politics.ox.ac.uk
    Title: digital news report
    Link: https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2025

  8. Source: reutersinstitute.politics.ox.ac.uk
    Title: dnr executive
    Link: https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2025/dnr-executive-summary

  9. Source: eu-digital-services-act.com
    Title: EU Digital Services Act Article 27, the Digital Services Act (DSA)
    Link: https://www.eu-digital-services-act.com/Digital_Services_Act_Article_27.html

  10. Source: newsroom.tiktok.com
    Title: how tiktok recommends videos for you
    Link: https://newsroom.tiktok.com/en-us/how-tiktok-recommends-videos-for-you

  11. Source: transparency.meta.com
    Title: ig feed recommendations
    Link: https://transparency.meta.com/features/explaining-ranking/ig-feed-recommendations/

  12. Source: transparency.meta.com
    Title: ig feed
    Link: https://transparency.meta.com/features/explaining-ranking/ig-feed/

  13. Source: transparency.meta.com
    Title: explaining ranking
    Link: https://transparency.meta.com/features/explaining-ranking/

  14. Source: reutersinstitute.politics.ox.ac.uk
    Title: Digital News Report 2025
    Link: https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2025-06/Digital_News-Report_2025.pdf

  15. Source: reutersinstitute.politics.ox.ac.uk
    Title: our podcast digital news report 2025 episode 2 ai and personalised news
    Link: https://reutersinstitute.politics.ox.ac.uk/news/our-podcast-digital-news-report-2025-episode-2-ai-and-personalised-news

  16. Source: research.google.com
    Link: https://research.google.com/pubs/archive/45530.pdf

  17. Source: github.com
    Link: https://github.com/xai-org/x-algorithm

  18. Source: youtube.com
    Link: https://www.youtube.com/watch?v=DQtJ2C2DM6I

  19. Source: youtube.com
    Link: https://www.youtube.com/watch?v=URI5GsOBznk

  20. Source: youtube.com
    Link: https://www.youtube.com/howyoutubeworks/recommendations/

  21. Source: youtube.com
    Link: https://www.youtube.com/watch?v=ZTWD3-K3r8w

  22. Source: shaped.ai
    Title: how youtubes algorithm works
    Link: https://www.shaped.ai/blog/how-youtubes-algorithm-works

  23. Source: solomonmg.github.io
    Title: twitter the algorithm
    Link: https://solomonmg.github.io/post/twitter-the-algorithm/

  24. Source: youtube.com
    Title: How Social Media Algorithms Actually Work (And How to Beat Them)
    Link: https://www.youtube.com/watch?v=8cQidXgtGmU
    Source snippet

    2 How Algorithms Shape What You See Online...

  25. Source: youtube.com
    Title: How Algorithms Shape What You See Online
    Link: https://www.youtube.com/watch?v=ioMx3tqI_VM
    Source snippet

    3 Algorithm Explained: How Social Media Decides What You See...

  26. Source: youtube.com
    Title: Algorithm Explained: How Social Media Decides What You See
    Link: https://www.youtube.com/watch?v=cqAwNuERNiw
    Source snippet

    4 How Algorithms Shape What You See Online...

  27. Source: youtube.com
    Title: How Algorithms Shape What You See Online
    Link: https://www.youtube.com/watch?v=m9umK2S4Q-g
    Source snippet

    5 An inside look at X's Community Notes | Keith Coleman & Jay Baxter...

  28. Source: youtube.com
    Title: An inside look at X’s Community Notes | Keith Coleman & Jay Baxter
    Link: https://www.youtube.com/watch?v=8dgyqYHLcCI

  29. Source: about.fb.com
    Title: About Facebook How AI Influences What You See on Facebook and Instagram
    Link: https://about.fb.com/news/2023/06/how-ai-ranks-content-on-facebook-and-instagram/

  30. Source: knightcolumbia.org
    Link: https://knightcolumbia.org/content/understanding-social-media-recommendation-algorithms

  31. Source: pewresearch.org
    Title: social media
    Link: https://www.pewresearch.org/internet/fact-sheet/social-media/

  32. Source: instagram.com
    Link: https://www.instagram.com/official/

  33. Source: about.instagram.com
    Title: ranking explained
    Link: https://about.instagram.com/blog/announcements/instagram-ranking-explained

  34. Source: instagram.com
    Title: DS3IKK8i Z8O
    Link: https://www.instagram.com/p/DS3IKK8iZ8O/

  35. Source: disinfocode.eu
    Link: https://disinfocode.eu/reports/instagram/5

Additional References

  1. Source: researchgate.net
    Link: https://www.researchgate.net/publication/401405020_Ranking_for_engagement_How_social_media_algorithms_fuel_misinformation_and_polarization

  2. Source: ofcom.org.uk
    Link: https://www.ofcom.org.uk/siteassets/resources/documents/research-and-data/online-research/other/pattrn_anayltics_intelligence_final_report.pdf?v=329852

  3. Source: brainforge.ai
    Link: https://www.brainforge.ai/blog/how-instagram-uses-ai-to-rank-your-feed

  4. Source: nexlev.io
    Link: https://www.nexlev.io/how-youtube-recommend-videos

  5. Source: kurdishstudies.net
    Link: https://kurdishstudies.net/menu-script/index.php/KS/article/download/3970/2684/7680

  6. Source: dictionary.com
    Link: https://www.dictionary.com/browse/official

  7. Source: merriam-webster.com
    Link: https://www.merriam-webster.com/dictionary/official

  8. Source: collinsdictionary.com
    Link: https://www.collinsdictionary.com/dictionary/english/official

  9. Source: ifo.de
    Link: https://www.ifo.de/en/cesifo/publications/2026/working-paper/ranking-engagement-how-social-media-algorithms-fuel-misinformation

  10. Source: esafety.gov.au
    Link: https://www.esafety.gov.au/sites/default/files/2022-12/Position%20statement%20-%20Recommender%20systems%20and%20algorithms.pdf

Topic Tree

Follow this branch

Parent topic

Think Before Sharing

Related pages 24

More on this topic 6