Educational companion dossier · Fact, interpretation, lived experience, clinical education, fiction, and mechanics are labeled separately. Scope & safety

AI PSYOPS TAXONOMY · CATEGORY 07

Algorithmic Perception Control

Ranking, recommendation, search, trending, moderation, notification, and distribution systems shape what people notice, regard as important, or treat as credible.

Primary level: Environment Documented systems · effects context-dependent Claim AIP-07-A01

CURRENT EVIDENCE ASSESSMENT

Ranking and recommendation systems structurally shape visibility; intentional AI-enabled perception control and downstream belief effects are context-dependent.

Stable claim identifierAIP-07-A01
Claim stageassessment
Currentness reviewed2026-07-27T00:00:00Z

DEPLOYMENT

Documented systems

Ranking, recommendation, search, moderation, and trending systems are deployed at population scale.

AUTONOMY

Environment-level optimization

Models continuously order information according to platform objectives and constraints.

PERSISTENCE

Structural and continuous

The influence environment persists as long as algorithmic mediation is active.

PROFILING ACCURACY

Behavioral prediction is common; inner-state inference is limited

Engagement histories can predict clicks, but do not reveal stable beliefs or emotions with certainty.

MEASURED EFFECT

Exposure effects documented; persuasion varies

Systems clearly alter what is seen; downstream attitude and behavior effects depend on users, content, and context.

Assessment basis

Assessment combines the exact owner-supplied category report with the bounded primary, official, platform, and peer-reviewed sources listed for this category. Dimensions are evaluated separately to prevent documented output from being mistaken for autonomy or effect.

What would change this assessment

Change the assessment when transparent audits link specific ranking interventions to replicated, durable outcomes across populations.

Prohibited inference

Do not infer strategic effect, universal deployment, or individual psychological state from this assessment.

A · DEFINITION

What this category means sources

Definition

Algorithmic perception control describes intentional or structurally induced shaping of what people encounter, notice, regard as popular, or treat as credible through ranking, recommendation, search, trending, moderation, and notifications.

Outside this category

All information systems must select and order content. Selection becomes a control concern when visibility is covertly distorted, adversarially manipulated, or optimized in ways that systematically undermine informed choice and representative understanding.

B · SIGNIFICANCE

Why it matters sources

The information environment mediates attention before any message can persuade. Visibility, repetition, and social metrics can make a topic appear important or widely accepted even when the underlying activity is artificial or unrepresentative.

C · CHANGE FROM PRE-AI PRACTICE

How AI changes the phenomenon sources

Machine-learning systems continuously predict engagement and personalize ranking at scale. Generative AI increases synthetic content and engagement signals, making popularity and authenticity harder to infer. Still, user agency and prior belief mean exposure does not translate uniformly into persuasion.

D · CAPABILITY STATUS

Separate evidence from projection sources

DOCUMENTED

Confirmed real-world use

Algorithmic ranking, recommendation, moderation, and search are pervasive; adversarial manipulation of trends and visibility is documented.

DEMONSTRATED

Demonstrated technical capability

Controlled studies show that ranking order and repeated exposure can influence attention and some judgments under bounded conditions.

EMERGING

Plausible near-term development

Generative content and synthetic engagement may intensify data pollution and feedback loops.

UNCERTAIN

Unsupported or unproven

A single feed or recommendation algorithm cannot be assumed to determine political identity or behavior uniformly.

E · KEY MECHANISMS

Conceptual mechanisms — not an operating procedure sources

Safety transformation: these descriptions identify system functions at a high level. Procedural steps, target criteria, scripts, evasion methods, and deployment workflows are intentionally excluded.

  1. Engagement-ranked feeds and out-of-network recommendations.
  2. Search ordering, data voids, and authority signals.
  3. Trending systems driven by volume and velocity.
  4. Downranking, demonetization, or visibility reduction without removal.
  5. Feedback loops between synthetic engagement, ranking, and perceived social proof.

F · EVIDENCE & EXAMPLES

What is known, measured, and still unknown sources

REACH IS NOT EFFECT. Publication, impressions, engagement, virality, or media attention do not by themselves establish persuasion or behavioral change.

Example 1 · Claim AIP-07-E01

Ephemeral astroturfing of trending systems

Platform-mechanism precedent documented; AI use not established in the study

What occurred
Researchers documented coordinated inauthentic activity that temporarily elevated topics in a trending system and was then removed.
What is confirmed
The attack pattern and manipulation of popularity signals were measured in the study.
Effect measured
The work establishes visibility manipulation, not AI involvement or downstream persuasion.
What remains unknown
The degree to which current AI agents use the same method, and any durable belief or behavior effects, remain unknown.
Source scope
The linked sources support the bounded statements shown here; they do not automatically establish intent, reach, persuasion, behavior, or strategic effect.
Correction trigger
Revise when a primary record, authoritative correction, adjudication, retraction, or stronger causal study changes the bounded statement.
Example 2 · Claim AIP-07-E02

Search and feed-ranking experiments

Exposure changes documented; generalized control claims unsupported

What occurred
Experimental and platform-supported research altered ranking, recommendations, or feed composition to examine exposure and attitude outcomes.
What is confirmed
Algorithmic ordering can materially change what participants encounter.
Effect measured
Some studies find attitude or knowledge effects while others find limited changes on selected outcomes.
What remains unknown
Results do not establish universal, covert population control or consistent long-term persuasion.
Source scope
The linked sources support the bounded statements shown here; they do not automatically establish intent, reach, persuasion, behavior, or strategic effect.
Correction trigger
Revise when a primary record, authoritative correction, adjudication, retraction, or stronger causal study changes the bounded statement.
Example 3 · Claim AIP-07-E03

Platform ranking and moderation changes

Internal optimization documented; causal effects mixed

What occurred
Large platforms changed feed ranking or visibility rules and researchers assessed exposure and polarization outcomes.
What is confirmed
Changes in what users saw were documented.
Effect measured
Some studies found substantial exposure differences but limited movement in core attitudes.
What remains unknown
Long-term effects and interactions with different platform cultures remain uncertain.
Source scope
The linked sources support the bounded statements shown here; they do not automatically establish intent, reach, persuasion, behavior, or strategic effect.
Correction trigger
Revise when a primary record, authoritative correction, adjudication, retraction, or stronger causal study changes the bounded statement.

G · RISKS & FAILURE MODES

Potential harms and reasons the capability may fail sources

Risks

  • Engagement objectives can reward outrage, repetition, or sensationalism.
  • Opaque downranking can create chilling effects and self-censorship.
  • Synthetic engagement can manufacture popularity and authority.
  • Researchers may lack the data needed to audit effects independently.
  • Automated moderation can disproportionately affect marginalized language or communities.

Limitations and failure modes

  • Exposure effects are heterogeneous and interact with prior belief and social context.
  • Platform data access is incomplete and changes over time.
  • Ranking audits can be confounded by personalization and account history.
  • Visibility change is not equivalent to censorship, persuasion, or behavior change.

H · DETECTION & DEFENSIVE INDICATORS

Signals for investigation, not automatic verdicts sources

Indicator rule: unless the source report supports a stronger conclusion, each signal below is suggestive rather than conclusive. Multiple independent signals and contextual evidence are required.

  • Sudden visibility spikes driven by short-lived or low-authenticity activity may be suggestive of manipulation.
  • Large gaps between visible engagement and persistent discussion warrant network analysis.
  • Abrupt reach changes without explanation may justify an appeal or audit but do not prove intentional suppression.
  • Search or feed results should be sampled across contexts before drawing conclusions about systematic bias.

I · GOVERNANCE & SAFEGUARDS

Accountability, transparency, and human protection sources

Provide meaningful user controls over ranking, chronology, and recommendation intensity.

Publish clear explanations of major ranking and moderation changes.

Enable independent researcher access with privacy-preserving safeguards.

Audit synthetic engagement, demographic impact, and false positives.

Reduce the use of raw engagement as an unquestioned proxy for public importance or credibility.

J · RESEARCH GAPS

Questions the evidence does not yet close sources

  • Causal effects of recommender systems outside short experimental windows.
  • Auditing methods that remain valid under personalization and frequent product change.
  • How synthetic engagement alters social proof and institutional trust.
  • How to provide transparency without enabling manipulation or compromising privacy.
REAL-WORLD INTERPRETIVE

L · SOURCES & REVIEW STATUS

Exact owner report, claim register, and reviewed sources

  1. Algorithmic Perception Control Owner-supplied report: Algorithmic Perception Control Report.md · 64,860 bytes · SHA-256 050bbe17757676e81641a6fe6d31b5959aa2b4966289f4020d632cd6d181776a

    Owner-supplied interdisciplinary research synthesis; exact source preserved in protected durable memory. External specialist review remains pending.

Claim-specific reviewed sources

  1. IEEE European Symposium on Security and Privacy · 2021-09-06 · Primary research

    Supports
    Documents coordinated manipulation of trending mechanisms through temporary inauthentic activity.
    Does not establish
    Does not establish AI involvement in the observed attacks and should be used as a platform-mechanism precedent, not as proof of AI deployment.
    Review
    LOCATED_AND_REVIEWED_AT_CITATION_LEVEL · Currentness checked for the bounded claim scope on 2026-07-27.
  2. SRC-27-NYHAN-META-ELECTION-STUDIESMeta 2020 U.S. Election Research Studies

    Meta and independent academic partners · 2023-07-27 · Mixed first-party data and independent research

    Supports
    Reports experimental and observational findings on feed changes, exposure, polarization, and political attitudes during the 2020 U.S. election.
    Does not establish
    Platform-specific findings do not establish that ranking has no effects in other settings; access constraints and design choices remain relevant.
    Review
    LOCATED_AND_REVIEWED_AT_CITATION_LEVEL · Currentness checked for the bounded claim scope on 2026-07-27.
  3. Meta · 2023-08-29 · Authoritative first-party platform disclosure

    Supports
    Documents the removal and analysis of coordinated inauthentic behavior, including use of GAN-generated profile imagery and the Spamouflage network.
    Does not establish
    Platform findings do not establish complete cross-platform activity, target exposure, or behavioral effect.
    Review
    LOCATED_AND_REVIEWED_AT_CITATION_LEVEL · Currentness checked for the bounded claim scope on 2026-07-27.
Selected works identified by the owner-supplied report
  • Geiger et al., research on ephemeral astroturfing and trending-system vulnerabilities.
  • Nyhan et al., Like-Minded Sources on Facebook Are Prevalent but Not Polarizing.
  • Epstein and Robertson, The Search Engine Manipulation Effect.
  • Bekavac and Mayer, auditing researcher data access under the Digital Services Act.

Exact source preservation and editorial currentness review do not constitute specialist certification, adjudication, legal advice, clinical review, or proof that every owner-report citation is current. Corrections remain open.

Page complete Algorithmic Perception Control Page label: CONTEMPORARY / ONGOING CLAIM — NOT SETTLED HISTORY