Absolutely yes. Even when you work with AI or research on it, you need to prove everything you do with statistics or the issue will be superfluous or have poor justification.
For my recent manuscript on machine learning in education and marketing, reviewers noted the lack of statistical validation:
- ROC-AUC and F1 scores are reported, but confidence intervals or statistical tests comparing models are missing. Provide more thorough statistical validation to strengthen claims.
- Perform paired t-tests or McNemar’s tests to validate performance differences between models. Report p-values with Bonferroni correction.
I don't have an MS in Statistics, but I'm now considering it.
Yes, with the rise of artificial intelligence, it is still worthwhile to pursue a master's degree in statistics because AI cannot perform advanced-level data analysis. It can be a data analyst, not a data scientist. Moreover, it is not good at mathematics.