What is the Trust Standard Protocol?
The open cryptographic standard that makes AI output verifiable — without having to ask the AI provider.
TSP is an open cryptographic standard that makes AI output verifiable. Every AI answer is signed so it can be verified mathematically — without having to ask the AI provider.
Think of TSP as "SSL for AI answers": an independent receipt that proves the answer is authentic, that it is grounded in documented sources, and that it has not been altered after signing.
The standard is MIT-licensed, open and can be implemented by anyone — from startups to large agencies. LexiCo AS in Tønsberg, Norway has built and maintains the spec, but no one owns the protocol.
Concrete examples
Three situations where TSP makes the difference between undocumented AI use and verifiable proof.
AI advisory in banking
How does the regulator know the AI advice was based on the correct regulations?
The bank cannot just say "the AI said so" — they must show which rules and documents the decision was based on. TSP signs every AI answer with a hash chain linking it to the sources, so the regulator can verify that the decision was correctly grounded.
AI diagnostics in healthcare
How does the hospital document that the AI diagnosis was based on up-to-date guidelines?
The hospital must be able to document that the AI assessment used approved clinical guidelines — not outdated or incorrect information. TSP provides a cryptographic receipt showing exactly which sources the AI drew from, signed at the same moment the answer was given.
AI decisions in municipalities
How can the citizen verify that the AI decision from the municipality was correct?
The municipality must be able to show the citizen — and the ombudsman — that the decision was based on the correct facts and laws. TSP creates an auditable link from decision to sources, so both citizen and supervisor can verify that the AI did not fabricate facts.
Back to TSP overview.