SOURCE AI_Liability_and_Insurable_Risk.png
MD5 260f579e · made 2026-09-16 · NotebookLM infographic · OCR tesseract 5.5.1

The Al Liability Gap: Turning “Decidable Drift” into Insurable Risk
THE PROBLEM: THE “UNDECIDABLE” RISK GAP

STATISTIC: KEY_FINDING:
onc Move from “Good”
0
0% Insurability to “In-Lane”

a ad Supporting Detail: We
for Black Box Al don’t judge if Al is ethical,
Supporting Detail: Current insurers we measure if it stayed in
exclude Al risk because software- its geometric lane.

only audits cannot prove where an
agent drifted from its intent.

/

[I |1|] <—
WD

Tamper-Evident

Receipt
Sealed Ledger

PROCESS_STEP:
Hardware-Anchored Attestation

Supporting Detail: Receipts are
recomputable offline against pinned public
keys, requiring no trust in the vendor.

Businesses

DEFINITION: Self-Insure STATISTIC:

Rice’s Theorem & Catastrophic Risk S PURPOSE: COMPARING CLASSICAL VS. THETADRIVEN TRUST PHYSICS

Semantic Drift. ke = 0.003

Supporting Detail itis (Th U am CLASSICAL SOFTWARE EVALS THETADRIVEN TRUST PHYSICS
aaiinnoeel e Universa

decide a progr output cane Drift Constant) Og Veriteation Basis veneceleeee

“good” from within software. En OPTUS Geometry a

Supporting Detail: A physics-
based constant that allows

erecta caer A Decidability: : Decidability:
actuaries to calculate information Undecidable / Decidable
decay across silicon and neurons.

(Rice’s Theorem) (-14ms latency)

i Audit Path:
Audit Path: :
A\y- Software-only “Vibe” 5d Tamper-Evident a
3

Sealed Ledger

Insurers Refuse to Sign
for Unmeasurable
Behavior

AN Gemini Notebook