If you spend some time looking in detail at the arguments that string theory requires supersymmetry, you'll find that they are not watertight. (How could they be, since we still can't say/don't know precisely what string theory is?)
Basically, some string theorists argue that that the usual classification depends too strongly on choosing nearly trivial boundary conditions and backgrounds, and that weirder things ought to be allowed -- like type 0 strings, Liouville backgrounds, gravitational duals of randomly-chosen CFTs, miscellaneous higher spin gauge theories, bosonic string tachyon condensation, strings above Hagedorn temperature. The experts differ on how believable these arguments are, and in whether the various bizarre things you get this way really should be thought of as part of string theory.
What this means is that no one can quite answer your question. No one knows for sure that they canSUSY is required. No one knows for sure that they can'tit isn't. This is why it is a very good idea to a) not believe most things you read in the popular literature, and b) not believe most things you read in the scientific literature.