13-17-507. Discrepancies -- substitution of results -- examination of machines. (1) If a discrepancy exists between the vote-counting machine totals and the manual count totals, the random-sample audit results must serve as the definitive record for purposes of the canvass.

(2) If the random-sample audit results in a discrepancy of more than 0.5% of total ballots cast or five ballots, whichever is greater, and if the discrepancy is determined to be due to the vote-counting machine and not to administrative or user error:

(a) the vote-counting machine involved in the discrepancy in that county may not be used in another election until it has been examined and tested by a computer software expert in consultation with a voting system vendor and approved by the secretary of state; and

(b) at least three additional precincts within the county must be audited for the office or ballot issue in question. If the county has fewer than three additional precincts, all remaining precincts must be audited.

(3) If the audit of the additional precincts results in a discrepancy for those precincts of more than 0.5% of total ballots cast or five ballots, whichever is greater, and if the discrepancy is determined to be due to the vote-counting machine and not to administrative or user error, the vote-counting machine involved in the discrepancy in that county may not be used in another election until it has been examined and tested by a computer software expert in consultation with a voting system vendor and approved by the secretary of state.

(4) The results of the random-sample audit must be public.