How CRV Digital Uses AI

We use artificial intelligence only as a controlled support tool within our workflow. Final outputs remain human-led, source-based, and reviewed before anything is delivered to a client. Our policy is designed to protect accuracy, confidentiality, and transparency, particularly where work may be relied on in legal, expert-support, research, or design contexts.

Human Reviewed
Source Based
Confidentiality Protected

What clients can expect

Human responsibility stays in place. Every output remains the responsibility of CRV Digital and the project lead handling the work.
Final deliverables are reviewed by a person. AI can assist with workflow, but it does not sign off, interpret, or replace quality assurance.

What AI will not be used for

No invention of facts. AI is not used to guess, embellish, or reconstruct disputed or case-critical facts.
No misleading evidential content. AI is not used to fabricate or alter source evidence, or to create misleading enhancements of CCTV or photographs.
No undisclosed substitution for expert judgment. AI is not used to generate expert opinion or replace documented methodology, assumptions, or validation.
No unapproved case-specific generation. AI is not used to generate case-specific people, disputed objects, or event details where those features are material, unless expressly approved and fully validated.

Confidentiality and data protection

Confidential, privileged, personal, or case-sensitive information is not uploaded into public or unapproved AI tools. AI tools are only used where client obligations, confidentiality requirements, and data protection standards can be maintained.

Approved Tools Only
Client Obligations Maintained
Sensitive Use Escalated

Need the full policy?

This page summarises the client-relevant points of the internal AI Use Policy and is intended as an accessible overview. Use the buttons below to download the full policy or see our other policies.

When disclosure may be made

Where appropriate, CRV Digital may disclose that AI-assisted tools were used in creating part of an output, particularly where that use could be relevant to interpretation, reliability, or scrutiny.

Usually permitted Internal drafting support, generic non-case-specific assets, generic textures, and administrative assistance.
Approval and validation required Approximate asset generation or inferred but non-critical geometry used in client-facing work.
Prohibited or exceptional only AI-created case-specific people, disputed objects without clear evidential basis, or event sequences presented as fact.