Two questions, checked at different moments
Recording consent is confirmed agreement from every participant in a conversation before it's captured — a decision made once, at the start, entirely separate from the later, ongoing question of who may access the resulting recording afterward. A system focused only on securing the stored recording — encrypting it, restricting who can retrieve it — has addressed the second question thoroughly while potentially never having addressed the first at all.
This distinction matters because the two failures look different and are caught differently. A weak access-control policy on an existing recording is a problem that monitoring and permission reviews can surface. A recording made without proper consent has already happened by the time anyone reviews access to it — the failure occurred at capture, not at any later access.
Consent has to be a gate on the capture itself
Building consent into the system means gating the capture action itself — not merely storing evidence of consent alongside the recording after the fact, but requiring confirmation before recording begins, in a way that's genuinely part of the capture workflow rather than a separate step bolted on around it after the fact. A consent record created after the fact demonstrates good intentions; a consent check that actually blocks capture until satisfied is what makes the guarantee real.