I was taking the test pretending to be my character just as an attempt to understand them more, I have a list full of it. In fact, I do tons of personality quizzes as them for this purpose. I don't really understand psychology and the test might be inaccurate, but hey it's quite fun and challenging to put myself into them.
My main character KĂ´ra is an INFJ. I thought he'd be an INFP and it kind of bugged me out that he has same type with me (Having a character that has too much sameness with me feels disgusting. The fact I hate self-insert doesn't help). Moreover, the test results and MBTI shitposts I found along made me unsure about my characters' characterization; do they act like their MBTI results? What if the audience know their MBTI types and they feel it's inaccurate?
Then I started to think "There is in fact a lot more personality of my characters than their MBTI results, as so real people." Even people with same MBTI result can have two distinguishable personality, and the test itself can be unreliable. So, I placed their MBTI as their fun fact trivia and nothing more.
So yeah, using MBTI as a fun game to understand more of the characters as people? Yes. Building and characterizing them around their MBTI result as a base? Not really good, you'd likely end up on making them complying the mold and stereotype based on the type.
(This applies to all personality tests)