[setr]

Its just a reality. It doesn’t matter if you cheat, if cheating isn’t effective. It also doesn’t matter if you cheat, if cheating is assumed (testing is only useful as a relative measure: if you expect 50% of the test will be cheated, then the comparison starts at 50%, not 0.)

The problem occurs when theres a discrepancy between groups: group A cheats, and group B doesn’t.

But if both cheat, then nothing changes, unless the cheating is total (theres nothing left to measure!). A subset of questions become meaningless, but assuming you as a test-maker can work around that, its fine.

Alternatively, you could just call it grade inflation.