When brought in, one of the perceived disadvantages of the new AI codes were that they may lead to a decrease in the accuracy of sire recording. ICBF have done a piece of work on this based on dairy genotyping over the last 3 years to look at the error rates (see tables below). At this stage the new codes are not leading to any large increases in sire errors.
Sire error rate per year
| Year | Number of errors | Total sires recorded | Error rate |
| 2015 | 801 | 15,694 | 0.051039 |
| 2016 | 825 | 13,581 | 0.060747 |
| 2017 | 669 | 11,500 | 0.058174 |
Short vs long AI codes
| Code | Sire Errors | No Errors | ||
| Number | % | Number | % | |
| Short AI Code | 1,913 | 5.59 | 32,299 | 94.41 |
| Long AI Code | 382 | 5.82 | 6,181 | 94.18 |
| All | 2,295 | 5.63 | 38,480 | 94.37 |
Sire error rate short vs long AI code per year
| AI Code | Year | Sire Errors | No Errors | ||
| Number | % | Number | % | ||
| Short AI Code | 2015 | 801 | 5.12 | 14,855 | 94.88 |
| 2016 | 738 | 5.97 | 11,627 | 94.03 | |
| 2017 | 374 | 6.04 | 5,817 | 93.96 | |
| Long AI Code | 2015 | 0 | 0 | 38 | 100 |
| 2016 | 87 | 7.15 | 1,129 | 92.85 | |
| 2017 | 295 | 5.56 | 5,014 | 94.44 | |