Volume and calling patterns alone should be sufficient to identify spam calls in most cases.
Audio analysis (probably AI/LLM-based) might be invoked on some, but even that is probably overkill.
Better: an appeals / review process which more-or-less automatically grants an initial reverse judgement, though if complaints continue from additional sources, that decision is reversed.
The real goal of my variant of this proposal isn't to nail the call originators but the networks (carriers) which either serve as origin or transit this traffic.
Audio analysis (probably AI/LLM-based) might be invoked on some, but even that is probably overkill.
Better: an appeals / review process which more-or-less automatically grants an initial reverse judgement, though if complaints continue from additional sources, that decision is reversed.
The real goal of my variant of this proposal isn't to nail the call originators but the networks (carriers) which either serve as origin or transit this traffic.