a/b testing does not hurt SEO as long as you are mindful of a few guidelines. For example, if you make sure of no cloaking as In user A/B testing, people are divided into groups and shown various iterations of the same page.
Cookies ensure that users always arrive at the same version of the page when they return.
Search engine bots are separated as they access this page, but they cannot be cooked. They consequently receive different iterations of the same page on successive visits. They detect the non-visible version concealed by the user testing process, giving the appearance of cloaking.
Given the briefness of the test, some testing suppliers might advise you that this is not a problem, but doing so is dangerous.