Methodology

What this measures

Measures the reach-weighted emotional valence of public discourse about AI across a defined platform basket. This is NOT a measure of public opinion: people who post about AI are not a random sample of the population.

The formula

I_raw(t) = 50 · ( 1 + tanh( κ · [ α·ES(t) + β·VV(t) + γ·MV(t) ] ) )
I(t)     = λ · I_raw(t) + (1 − λ) · I(t−Δ)

ES = engagement-weighted mean valence, shrunk toward 0 on small samples
VV = seasonally-adjusted, robust-z-scored log volume growth
MV = outlet-prestige-weighted media alarm, normalised to trailing P95

alpha_emotional_severity

0.5

beta_volume_velocity

0.25

gamma_media_virality

0.25

kappa_gain

1.5

shrinkage_k

200

author_weight_cap

0.005

min_authors

150

Every component is normalised to [−1, 1] before the weighted sum. This matters: without it, the media term sits on a different scale from the others and silently dominates, making the weights decorative.

Where the data comes from

No single source exceeds 0.22. The index is defined over a basket, so a platform revoking access degrades precision rather than killing the measurement.

Bands

• 0–24 Extreme Trust
• 25–44 Trust
• 45–55 Neutral
• 56–74 Fear
• 75–100 Extreme Fear

0 is extreme trust and 100 is extreme fear. Note this is inverted relative to the Crypto Fear & Greed Index, which confuses people who know that one.

Known limitations

• —Discourse is not opinion. The sample over-represents technical, Anglophone, and highly-online populations.
• —Emotion classification accuracy is bounded; sarcasm is the dominant failure mode and biases the index toward under-reporting fear.
• —Confidence intervals reflect sampling uncertainty only, not model error.
• —The index is not a forecast and has no established predictive validity.

We publish these because the first thing a competent critic does is look for them. Stating the limitations before anyone else does is the only version of this that survives scrutiny.

About the demonstration data

When this deployment reports data_mode: synthetic, the underlying documents are generated rather than collected. The formula is the real one — the same code that runs against live data — but the corpus it consumes is invented.

Two consequences worth stating plainly. The numbers describe nothing about any real company, model or person. And the quotes are generated from templates with fictional handles: this platform does not fabricate statements and attribute them to real, named people, and running a demo is not an exemption from that.

Machine-readable

The full definition — parameters, basket, outlet weights, bands — is served at /v1/meta/methodology. Publishing it costs nothing: the hard part is the pipeline, not the arithmetic.