AI Fear & Trust Index

How frightened is the internet of AI?

A live 0–100 index of fear and trust in AI systems, computed from public discourse. Every number here ships with its sample size and its confidence interval — click any entity for the detail, or read the methodology.

Now

61.0Fear

current reading

21 Sept 11:05 UTC

Yesterday

67.9Fear

▼ 6.9 toward trust

20 Sept 09:45 UTC

Last week

52.0Neutral

▲ 9.0 toward fear

14 Sept 09:45 UTC

Last month

no reading this far back yet

Composite index

Mon, 21 Sep 2026 11:05:00 GMT
025507510061.0FEAR
24h ▲ 7.17d ▲ 6.795% CI 59.262.8polarisation 10

The split

Fear10/100 contestedTrust
95%5%

One side clearly dominates the conversation right now.

Last 14 days

Based on 213 documents from 115 distinct authors (effective sample 124)low_diversity

Sources: News RSS 51%, mastodon 25%, Hacker News 23%

Share on BlueskyShare on X

What this sample covers

140 of 213 posts are too new to judge by reach. They are counted at the lowest weight rather than dropped, because discarding every recent post would leave the six-hour window permanently empty. A reading made mostly of these is early, not settled.

    Doom Clock

    Full ranking

    8h 29m to midnight

    Doom Index 29.3 / 100

    This measures discourse, not risk. It reflects how much and how intensely people are talking about existential risk in relation to each system. It is not a prediction, a probability, or an assessment of actual danger.

    Which AI is feared most

    Index value over the last 24 hours, by dominant fear vector · Higher is more feared

    1 of 13 entities

    The organisations behind those systems. A company's reading comes from posts about the company itself — it is not a total of its products.

    0102030405060OpenAI: 51.851.8OpenAIneutral
    • OpenAI51.8n = 26Mixed — no single driver

    12 entities are held back this window for insufficient data, across both levels.

    See every entity as a card — one page with the full side-by-side table, the fear/trust split per entity and the sample behind each number.

    What kind of fear

    The headline number is one scalar; this is the diagnosis · Higher is more intense

    • Existential risk73.1▼ 1.7

      n = 202

    • Jobs & economy56.3▲ 0.9

      n = 18

    • Deepfakes & deception50.7▲ 0.0

      n = 8

    • Privacy & data50.3▼ 0.3

      n = 3

    What each entity is feared for

    Share of each entity's fear mass by vector · Taller segment means more of that fear

    • Existential risk
    • Jobs & economy
    • Deepfakes & deception
    • Privacy & data
    0%20%40%60%80%100%OpenAI · Existential risk: 41.7%42OpenAI · Jobs & economy: 23.7%24OpenAI · Deepfakes & deception: 14.8%15OpenAI · Privacy & data: 19.8%20OpenAI

    Fear against attention

    Index value versus how much is being said. Marker size is effective sample. · Up is more feared, right is more discussed

    Loud and feared50.55151.55252.55353.524.52525.52626.52727.5OpenAI — Documents in the window 26.0, Index value 51.8OpenAIDocuments in the windowIndex value

    Which AI does the internet fear most?

    Volume and intensity of existential-risk discourse, per entity · Higher means more doom talk — not more danger

    0 of 6 entities
    0102030all AI discourse

    All entities

    Sort any column. Sample size sits beside every score, because a 74 from 60 documents and a 74 from 2,000 are not the same claim.

    1 of 1 entities
    OpenAIOpenAIno dominant vector51.851.4%0.1%0.22621Detail →

    What is being said

    The most-amplified statements in the window, fear and trust alike.

    • Ten days that changed the AI race : How fears of rogue systems grew
      @GDELTExistential risk__global__estimated impact 98
    • If Scots led on AI , the world would be safer : AI scholar fears
      @GDELTExistential risk__global__estimated impact 98
    • Extra : A Race To Recklessness ? Tristan Harris Breaks Down AI Threat To Humanity
      @GDELTExistential risk__global__estimated impact 95
    • Amid growing AI fears, King Charles meets with industry leaders in Scotland. King Charles convenes leaders of artificial intelligence companies at a summit in Scotland, amid mounting fears that unchecked AI could threaten humanity.
      @News RSSExistential risk__global__estimated impact 99
    • AI's imminent hacking threat is hiding in plain sight. An unprecedented wave of AI-powered, human-driven cyberattacks is coming, security experts say — and it's a far more urgent risk than theoretical scenarios in which AI wipes out humanity. Why it matters: The great Septem
      @News RSSExistential risk__global__estimated impact 99
    • "I am the Hoax Buster": Trump's war on AI doomers gets personal. President Trump declared war on the AI safety panic Monday, dismissing the industry's apocalyptic warnings as part of a "sick conspiracy" to sabotage America and his legacy. Why it matters: Trump is rewriting
      @News RSSExistential risk__global__estimated impact 99

    Emotional scores are machine-generated estimates about a statement, not an assessment of the speaker.

    Key definitions

    Frequently asked questions