Independent assurance · Cyber security + AI

Know what you're trusting.

AI Trust Scan for vendors, products and organisations.

AI Trust Scan · published assessment24 Aug 2026
Anthropicanthropic.com

ref vas_d12bf8fc9cd7003fbd

17 public sources gathered
20 claims extracted and classified
Certifications checked against registries

TRUSTYCYBER Score

64/ 100
C

Approve with conditions

Public evidence · point-in-time · not an assurance opinion

View the live report

How TRUSTYCYBER works

Five steps from claim to assurance.

01Claim02Evidence03Assessment04Finding05Assurance
“We don’t train on your data.”Privacy statement · retrieved 24 Aug 2026Customer data · 15% of the scorePartially supportedIndependent review, signed

AI Trust Scan

Choose what to assess.

Who are you trusting?

The organisation behind an AI service - its governance, security foundation, supply chain and independent assurance.

Run an AI Trust Scan

Free · results within a day · no documents requested

The public scan is free. When you need evidence examined, tested and signed, commission an assessment →

Published assessment · Vendor24 Aug 2026

Anthropic

anthropic.com · 20 claims examined · 17 public sources

ref vas_d12bf8fc9cd7003fbd

Approve with conditions

TRUSTYCYBER Score

64 / 100
C

Assurance dimensions · weighted components of one score

Organisation & AI governance15%60
AI system15%47
Model10%60
Customer data15%60
AI supply chain10%40
Security foundation15%85
Independent assurance evidence10%100
Legal & contractual10%60

Evidence coverage 50% · score ≠ decision · methodology versioned

Public evidence · point-in-time · not an assurance opinion

Methodology View the live report

The Trust Directory

Every report can be checked.

Every published AI Trust Scan is listed in the Trust Directory and resolves to a full, source-linked report that anyone can check.

Browse the Trust Directory

How far do you need to go?

Three levels of assurance.

The free scan covers public evidence. A commissioned report goes deeper on the same outside-in basis. An inside-out assessment examines internal evidence with your cooperation.

01 · Free

AI Trust Scan

Automated · public evidence · point-in-time

The scan on this page. Public evidence gathered, certifications checked against official registries, and a graded TRUSTYCYBER Score with source-linked findings, published to the Trust Directory.

02 · Commissioned

AI Trust Report

Outside-in · human-verified · quality-reviewed

A current, in-depth assessment of a vendor, product or organisation. Certifications are verified by a person against official registries and held documents, and the report is reviewed before release. Reading an existing report costs less than commissioning a new one.

From $390 · new reports from $1,800

Request a report

03 · Engaged

Independent assessment

Human-led · scoped · signed

A qualified auditor examines and tests the evidence: a supplier reviewed for your decision, or your own organisation assessed inside-out with your cooperation. Findings are determined and signed. Covers vendor reviews, solution assurance and ISO/IEC 27001 & 42001 internal audits.

Scoped and quoted at fit check

Check your fit

Report prices are confirmed before any work begins or payment is taken. The scan and reports assess public evidence and are not assurance opinions. An independent assessment provides one.

Independence

Engagements are led and signed by Andrew Robinson, with more than two decades in cyber security and assurance. Conflicts and prior involvement are assessed before any engagement is accepted, and work is never represented as independent where objectivity cannot be maintained.

ISO/IEC 27001 Lead Auditor
ISO/IEC 42001 Lead Auditor
JASANZ Technical Expert & Assessor
ASD IRAP Assessor

Know what you're relying on.

Run a free AI Trust Scan now, or commission independent assurance when the decision matters.