An A/A test is split-test that pits two identical versions against each other. Visually, it looks like this:

Version 1 and Version 2 are identical.

A/A tests are performed to validate that you’ve set-up and run the test properly — and that the data coming back is clean, accurate, and reliable.

A/A tests help rule out false positives in your data. While some may argue they're not necessary to perform, they are considered a testing best practice.