Short answer: categories are much less important in Minimalism, as long as all of the relevant features check. Some even try to avoid using labels at all, to the extent that is possible (but as discussed recently by Chomsky and others, and that leads to other problems).
Your observation/objection makes sense to me, although it matters less if you also think that (now in Minimalism) categories matter less.
One interesting point would be that it's no longer clear what "relative clause" would mean, except to say essentially that the structure of these two constructions is not substantially distinct. Labeling things like "relative clauses" would make more sense in earlier Generativism (e.g., rewrite rules) or Construction Grammar than trimmed down Minimalist theories. To be clear, Minimalism is a family of theories all sharing one property: they're trying to do away with as much of the complex earlier theories as possible, and this usually proceeds by individual authors addressing a single point (or several related points) in a paper, without actually proposing a fully theory, but rather just testing what happens when you remove something. In that sense, this idea is perfectly in line with Minimalist reasoning: what happens if we start re-using "relative clause" structure for other things-- how well does it work? How much (or little) can we get away with? The result is much less focus on distinctions between constructions (as in Construction Grammar, etc.) and more (perhaps overly broad) generalizations. So in short the only way to answer this is to carefully read the paper(s) that propose it and see how it works with the other assumptions that they retain in their particular Minimal theory. In principle there's nothing wrong with considering RCs and other similar constructions to have the same structure-- if it works.
