Re: Standard biometric flaw
Can you just use a photo of your iris? Because that's really not secure at all.
I think it's too early to tell, but in my opinion you're looking at a clever volume test of new technology that Fujitsu is developing, smartphones are a really quick way to do a mass rollout of something that is still subject to improvement. In case you didn't know, Fujitsi also develops sensors for palm recognition, and how these work may give a clue as to why eye recognition may actually work.
For a start, these are depth readers, so they look "beyond" your skin for vein patterns, and a picture won't do. Next, they had to simplify analytics already as the original ones produced so much data that a pass/fail took seconds (if I recall correctly the first ones took well over 10 seconds) so they may have found a new balance between resolution and security and may have ported all that learning to this phone and iris scanning.
I'm now entering the realm of speculation, but I think it's plausible to assume that this eye scanner may look for vein patterns instead of iris matrix. They may swell up after a night out, but AFAIK the pattern doesn't change (anyone with a medical background? Is this correct?). Alternatively, few are focusing on iris recognition of late, so Fujitsu may have come up with something new.
As for how to use that, there are already various deployment models out there that don't require your biometrics to travel off the device - you'd just use a locally stored hash of the biometrics to open a credentials strongbox in the phone (which is where all the more traditional challenges hide :) ).
So, based on past performance, I reckon this may indeed be interesting enough to keep an eye on, so to speak :)