Cognitive Impact · Harm Index

The platforms that measurably alter your mind, ranked by what can actually be proven

Below is what the peer-reviewed research establishes about how a platform changes your attention, your memory, your mood and your read on other people — and, in the same frame, who owns it. Two things to know before you start. Every claim carries a confidence tier, so where the evidence is thin it says so in the same size type as everything else. And the failed replications sit in the same list as the successful ones; a ledger that hides its null results cannot be checked.

11subjects

6mechanisms

4ownership

56primary-tier

COGNITIVE HARM INDEX 0–100 · documented harms × severity × reach · full leaderboard ↓

Platform Impact Profiles

One profile per platform: what the peer-reviewed research establishes about its effect on cognition and mental health, set beside the entity that owns or controls it.

AI System

ChatGPT

42 · Moderate

Conversational AI as cognitive prosthesis.

2 documented harms · OpenAI

Crypto / Stablecoin

Tether (USDT)

25 · Documented

Largest stablecoin; financial gambling rails.

1 documented harms · Tether

Professional Network

37 · Documented

Professional identity and hiring network.

2 documented harms · LinkedIn

Social Network

Facebook

55 · Moderate

The original engagement-maximising social network.

4 documented harms · Meta Platforms

Instagram

44 · Moderate

Image-first social comparison engine.

3 documented harms · Meta Platforms

Snapchat

33 · Documented

Ephemeral messaging and streaks for adolescents.

2 documented harms · Snap Inc.

X (formerly Twitter)

36 · Documented

High-velocity public discourse platform.

2 documented harms · X Corp

Surveillance Infrastructure

Palantir Foundry / Gotham

29 · Documented

Pentagon-grade analytical surveillance.

1 documented harms · Palantir Technologies

Video Platform

Rumble

40 · Moderate

Alt-tech video platform.

1 documented harms · Rumble

TikTok

69 · High

Short-form video, algorithmically auto-served.

4 documented harms · ByteDance

YouTube

53 · Moderate

Recommendation-driven long-form video.

2 documented harms · Alphabet

Cognitive Harm Footprint Leaderboard

A transparent 0–100 score per platform combining documented harm count, severity weighting (suicide/addiction/radicalisation rank highest), reach, and the presence of engineered mechanisms. Public methodology write-up is in progress; in the meantime each platform page shows its own component scores.

Consumer platforms — per-user engagement harms

TikTok ByteDance High

4 documented harms · severity 22 · reach 22

Facebook Meta Platforms Moderate

4 documented harms · severity 20 · reach 10

YouTube Alphabet Moderate

2 documented harms · severity 16 · reach 22

Instagram Meta Platforms Moderate

3 documented harms · severity 14 · reach 10

ChatGPT OpenAI Moderate

2 documented harms · severity 9 · reach 18

Rumble Rumble Moderate

1 documented harm · severity 8 · reach 22

LinkedIn LinkedIn Documented

2 documented harms · severity 10 · reach 12

X (formerly Twitter) X Corp Documented

2 documented harms · severity 11 · reach 10

Snapchat Snap Inc. Documented

2 documented harms · severity 8 · reach 10

Tether (USDT) Tether Documented

1 documented harm · severity 5 · reach 10

Surveillance infrastructure — population-level harms (distinct pathway: chilling effect, involuntary exposure, behavioural modification under observation)

Palantir Foundry / Gotham Palantir Technologies Documented

1 documented harm · severity 7 · reach 12

Note: these scores are not comparable to the consumer-platform scores above, and combining them would be a category error. The harm is borne by the people a system observes, not by anyone who chose to open an app. Reach here counts who can be watched — not who signed up.

CASEFILE Nº 1 DR. DUH · DEVIL'S ADVOCATE Speculative · graded end to end · argues against itself

Eighteen Months

The sequence by which an ordinary person is separated from their own judgement, one documented mechanism at a time. Every mechanism is cited. The chaining is conjecture and says so. Part two argues against part one at equal length.

Documented / strong Moderate Contested Extrapolation Read the casefile →

Mechanisms of Harm

The specific UI/UX design patterns that drive these outcomes. These are deliberate engineering choices that erode user agency by design.

🚫

Frictionless Design

Removing every pause, prompt or step that might let the user think.

3 documented effects

∞

Infinite Scroll

A virtual slot machine: variable reward, no stopping cue.

4 documented effects

▶

Auto-Play

Removing the choice to start the next piece of content.

3 documented effects

⚙

Algorithmic Recommendation

The feed chooses what the user thinks about next.

2 documented effects

🔔

Engineered Notifications

Manufactured interruptions that retrain attention.

2 documented effects

🔥

Streaks & Engagement Locks

Loss-aversion turned against the user.

2 documented effects

Accountability Map

Documented ownership links between Closed-Loop individuals and the platforms whose harms are catalogued here. Each link follows from public records to peer-reviewed harm.

Entity	Platform	Role	What the public loses	Loop
Alex Karp	Palantir Foundry / Gotham	CEO and co-founder of Palantir Technologies	Operates the infrastructure for Pentagon cognitive-warfare initiatives designed to make populations legible, predictable and governable.	View Loop →
Howard Lutnick	Tether (USDT)	Family-trust counterparty (Dynasty Trust A loan); Twenty One Capital partner	Profits from a stablecoin underpinning crypto trading documented to produce gambling-pattern cognitive distortions; conflicts intensify the cognitive harm.	View Loop →
Peter Thiel	Palantir Foundry / Gotham	Co-founder, chairman and largest shareholder of Palantir	Owns the company providing operational backbone for cognitive-warfare doctrine and mass surveillance.	View Loop →
Peter Thiel	Rumble	Investor via Narya Capital and personally	Funds a platform documented to host conspiracy and extremist content at high density, degrading the information environment for civic decisions.	View Loop →

Syndrome Cluster Pages

Each cognitive domain has its own evidence story — peer-reviewed studies, replications, contradicting findings, and a compounding-effects panel listing every platform whose published research touches it. Click a tile to read the full chain.

Adolescent Mental Health

Documented changes to anxiety, depression, self-harm, and identity formation in users aged 10–19 — restricted to studies where platform exposure is the variable actually tested, not assumed.

19 studies 16 peer-reviewed 5 hypothesis-flagged

Attention & Focus

Measured impact of interruption, fragmentation, and reward-on-a-variable-schedule design on sustained attention, working memory, and task-switching cost. The effect sizes are smaller than commonly claimed.

6 studies 5 peer-reviewed

AI-Induced Psychosis (hypothesis)

A hypothesis-level cluster of clinical reports describing dissociative, paranoid, or grandiose ideation after heavy chatbot use. Not a recognised diagnosis. Every entry is caveat-flagged, and the obvious rival explanation — that people already deteriorating gravitate to heavy chatbot use — has not been ruled out.

8 studies 7 peer-reviewed 5 hypothesis-flagged

Sleep Disruption

Effects of late-night use, blue-light exposure, and infinite-scroll engagement loops on sleep onset latency, total sleep time, and sleep architecture.

4 studies 3 peer-reviewed

Compulsive Use & Behavioural Addiction

Studies measuring loss-of-control, withdrawal, and tolerance markers attributable to platform-engineered variable-reward systems.

5 studies 5 peer-reviewed

Social Comparison & Body Image

Measured effects of curated, filtered, algorithmically-amplified peer content on self-evaluation, body satisfaction, and disordered eating risk.

6 studies 4 peer-reviewed

Aspirational Pacification (theory)

The claim that influencer culture absorbs labor anxieties into individualized aspiration — hustle content, lifestyle consumption, manifestation — dampening collective response to structural conditions. The mechanism's parts are separately documented in peer-reviewed work: aspirational and hope labor shift economic risk onto the individual, system-justifying belief measurably suppresses protest, and manifestation belief substitutes felt success for the real thing. The composite claim — that this prevents systemic rebellion — is a theory and is flagged as one: no study measures influencer exposure against collective action directly.

7 studies 4 peer-reviewed 2 hypothesis-flagged

Studies Evidence Ledger

A DOI-first ledger of the peer-reviewed studies, and the replication attempts, behind everything above. Hypothesis-level diagnostic labels — "AI-induced psychosis" and its relatives — carry the standard caveat: they are working descriptions of an observed cluster, not recognised diagnoses.

68 studies on file 56 peer-reviewed primary 24 replication attempts (4 failed) 12 hypothesis-level labels (caveat-enforced)

primarycompulsive_use2025

The Effect of Deactivating Facebook and Instagram on Users' Emotional State

American Economic Journal: Economic Policy (NBER WP 33697)

Single-paper claim · pending independent replication

secondaryexecutive_function2025

The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers

Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems

Single-paper claim · pending independent replication

primaryexecutive_function2025

Generative AI without guardrails can harm learning: Evidence from high school mathematics

Proceedings of the National Academy of Sciences

1 replication attempt on file

primaryai_psychosis2025

Delusional Experiences Emerging From AI Chatbot Interactions or “AI Psychosis”

JMIR Mental Health

ai_induced_psychosis (hypothesis-level label; not a recognised clinical diagnosis) — Authors describe 'AI psychosis' explicitly as an emerging clinical phenomenon under investigation, not an established diagnosis. The viewpoint synthesizes scattered case reports and clinical commentary; it does not establish causation. Cited evidence includes Canadian Broadcasting Corporation reports of individual cases and peer-reviewed commentary by Østergaard. Authors note that chatbot dynamics may reinforce mania-like symptoms (elevated mood, grandiosity, impulsivity) in addition to delusions, expanding the proposed phenomenology beyond psychosis narrowly defined. The label remains a working construct; not in DSM-5-TR or ICD-11.

Single-paper claim · pending independent replication

primaryai_psychosis2025

Generative Artificial Intelligence Chatbots and Delusions: From Guesswork to Emerging Cases

Acta Psychiatrica Scandinavica

ai_induced_psychosis (hypothesis-level label; not a recognised clinical diagnosis) — Author explicitly frames this as transitional evidence: title is 'From Guesswork to Emerging Cases', acknowledging the original 2023 hypothesis was 'merely based on guesswork.' Author calls for 'empirical, systematic research' rather than asserting causation. Most underlying cases are media-documented or anecdotal correspondence, not clinically verified primary research. The 'ai_induced_psychosis' construct remains a working hypothesis; it is NOT a DSM-5-TR or ICD-11 diagnosis. Causation between chatbot use and delusional ideation remains undemonstrated by controlled study.

Single-paper claim · pending independent replication

primaryai_psychosis2025

I tweet, therefore I am: a systematic review on social media use and disorders of the social brain

BMC Psychiatry

digital_schizophrenia (hypothesis-level label; not a recognised clinical diagnosis) — Authors explicitly frame 'Delusion Amplification by Social Media' as a new conceptual model, not a clinical diagnosis. They write that increased social media usage 'may be associated with psychotic spectrum phenotypes, especially delusionality, by the decoupling of inter and intra-corporeal cues integral to shared reality testing' — framed as a proposed mechanism warranting empirical testing, not an established disorder. The review is observational-correlational; causation is not established. 'Digital schizophrenia' does not appear as an author term; the closest construct offered is sub-clinical psychotic-spectrum amplification via the proposed model.

Single-paper claim · pending independent replication

Why this matters

When an owner says his platform connects people, this page answers with the citations — the documented anxiety, the documented attention cost, the documented adolescent brain changes. Nothing here is invented. It is the established literature, tied to the people who profit from the design choices that produce it.