When to use Bonferroni

I am working with a mark-all-that-apply survey question that has 8 categorical response options. I am comparing responses statistically between two groups (dependent variables).

1. Are Bonferroni corrections supposed to be made only when you are comparing multiple groups, or does the number of response options affect whether you should run a correction (ex., would you do it for 3 groups and 10 response options and also for 1 group and 10 responses options)?

2. Is a Bonferroni correction necessary to run when you are comparing just two groups?

3. Since each response on a mark-all-that-apply could be thought of as a separate yes/no question, and the response options are independent of each other, is it necessary to run a correction on the t-tests I run for comparison between groups for each response, or is this only needed for a single-response questions?

Thank you for any help!