The Zemi Method: An Evidentiary Methodology for Investigation and Assurance, Version 1.0

The Zemi Method is a published, versioned evidentiary methodology governing digital forensic investigation and AI assurance. Its premise is that every claim relevant to a decision, including claims produced by an organization's own systems, is a hypothesis until evidence from an independent source supports it. The method defines five governing principles, a six-phase investigative lifecycle with a mandatory human adjudication gate, a categorical classification of findings as known, assumed, or undetermined, and a standing defensibility check. Its fourth principle, AI-assisted, never AI-decided, permits automated systems to retrieve, correlate, summarize, propose patterns, and draft, while barring them from conclusions, adjudication, and sign-off. The same evidentiary standard extends to the assurance of AI systems, where the system itself becomes the subject under examination. This record contains Version 1.0 of the method and its companion document, Foundations and Positioning of the Zemi Method, which identifies the traditions the method draws on, acknowledges its predecessors, and bounds precisely what is and is not claimed as original. The method is maintained in dated, versioned form. Canonical home: https://zemimethod.com. SHA-256 integrity hashes of the published files are recorded at the canonical home and in the method's version repository. zemi-method-v1.md: 73b5ab638bea58bfc50407e2f132a8954687f4155711fe541a074b0d109f8a37zemi-method-foundations-companion-v1.md: 9a73facc72b2b2107b443341c81f7eaa25388177eaf849f0743d013744f455ea © 2026 Kevin V. Watson. All rights reserved.