It depends how bad they are. I'd say the threshold between bad and useless for me is: 1. Do you have to change unrelated tests to still pass as you add features? 2. Do the tests ever catch issues as you modify related code?
They may be bad, but if they work and don't get in the way, there's no reason to remove them. Even if they're useless, you're pretty much guaranteed to prove yourself incorrect the moment you delete them (you'll introduce an issue they would catch). In most cases it's better to just leave stuff in until it can be rewritten properly - removing should be the last resort.
