Firstly, you're comparing chalk and cheese: or rather, cheddar and cheese. RDF is XML, so you're asking "what's the advantage of cheese over cheddar?"
Secondly, without knowing something about your problem one can't possibly advise on the merits of using different technologies in the solution.
In choosing an XML vocabulary (and RDF is just an XML vocabulary) there are two questions you should ask: (a) does this vocabulary provide a good model of my information space, and (b) are there useful applications that can process this XML representation of the information?
In the case of RDF, my personal view is that you might get the answer yes to both questions if you are capturing metadata about resources on the web, and you will probably get the answer no if you are capturing anything else - though the RDF enthusiasts would tell you to use RDF to process your company's payroll, I would not.
Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference