About
Why AuditTrail exists.
The problem
The legal profession is using generative AI in client matters at scale, but there is no built-in record of that use. A practitioner who relies on AI today has no way to demonstrate, months later in court or to a regulator, what was asked, what was returned, what was relied on, and whether anything has been silently altered since.
The cases
In Hussein v. Canada, 2025 FC 1060 (with costs at 2025 FC 1138), the Federal Court ordered $100 in personal costs against counsel for failing to disclose the use of Visto.ai in immigration submissions. The Chief Justice of the Federal Court has issued a formal notice (May 7, 2024) requiring AI disclosure in materials submitted to the Court. The Law Society of British Columbia, the Law Society of Alberta, and the Lawyers Indemnity Fund of British Columbia have issued parallel guidance.
The solution
Every AI conversation captured by AuditTrail is cryptographically hashed and linked into a SHA-256 hash chain. Any edit to any record breaks the chain at the precise position of the alteration. The proof is the chain itself; it can be verified by anyone — opposing counsel, a regulator, a court — without trusting AuditTrail’s servers.
What AuditTrail offers
A browser extension that captures conversations with Claude, ChatGPT, and Gemini. A dashboard for review and export. A public verifier at audittrail.ca/verify. An exportable audit packet (PDF + JSON chain) for filings, disciplinary inquiries, or insurance claims.
Scope
AuditTrail is voluntary infrastructure. Nothing in the tool prevents a practitioner from uninstalling the extension, using a different account, or capturing AI activity outside the system. The value is not enforcement; it is defensibility. A practitioner who installs AuditTrail can demonstrate, with cryptographic proof, what was used and what was relied on. A practitioner who does not has no such proof if later asked. The choice belongs to the practitioner — and to the firm policy that governs them.