Not just accurate, but easily verifiable through observation. I wonder about this change in Georgia - if the recruited counters are not impartial then there could be co-ordinated bias (e.g. applying different standards to rejecting Harris / Trump ballot papers or introducing deliberate errors).