In theory, this isn't actually that out-there as an addition to a treatment regimen, although the trial should be an order of magnitude larger to produce meaningful data. What we'd hope for is a means of giving the patient a quantifiable, self-directed method of practicing certain aspects of his or her cognitive behavioral therapy -- there's a lot more to therapy than what takes place at the therapist's office. The danger comes from a product that allows the patient to learn to beat the game, rather than improving his or her skills in the real world. (This is where so-called "brain training" games for general entertainment have failed: Play memory cards for a few hours a day, and you'll get very good at turning over memory cards. You still won't be able to find your keys in the morning though.)
Schizophrenia basically means that a person has difficulty assigning priority to ideas. The toast you actually just put in the toaster has no more significance than the goofy idea that just popped into your head about your ex. Sounds reasonable until you consider thinking that way nearly all the time, and actually trying to get anything done. Add a dash of natural human paranoia, and it can cause some serious harm.
We'll hope for the best, but I still prefer to see any new treatment given the level of scrutiny we instinctively give to a new (molecular) medication.