Does learning ML / AI require one to be good at Competitive coding?

To learn it you only need to read about it and have an understanding of statistics and calculus, depending on what type of ML. Why would you need competetive coding?

No.

No, it's statistics that you should focus on learning.

Competitive coding is a sport. Even some good competitive coders aren't good devs. Not only statistics but you should do some hands on deployment of ML models, integrate your models with applications. Also learn some data viz tools and excel. It makes your life easier.

No. Nothing in the industry requires competitive coding.

Learning ML/AI really requires a good understanding of a programming language ( e.g. Python) , some understanding of data engineering, a little bit of statistics, linear algebra, and some basic calculus. None of this requires competitive level programming.

Absolutely nope lmao.
You don't need to be good or practising or let alone being into competitive programming to get into ML/AI, and having studied ML; I can tell that it will help you much less than traditional programming, where competitive programming isn't that beneficial unless you're able to discern when to write clean, readable and maintainable code (basically any coding related task aside from Code Golf and Competitive Programming) and when you don't have to worry about those.

Competitive Programming is a Sport, It requires programming skills plus ability to solve problems fast, For ML, or any other CS Field, Programming is required, but not at the pace at which competitive Programming demands.