umm...it's not possible to "create credit BitCoin out of thin air".  One must consume significant processing power to find a (previously undiscovered) large number whose SHA-256 hash is below the current target.

Now of course private individuals are free to set up their own private fractional reserve bitcoin banks, which could offer on-demand deposits holding people's bitcoins with a certain interest rate and give out loans at another interest rate (and keep the difference as profit), while only holding a certain percentage of total assets.  But then that bank would be at risk of default if all of its creditors simultaneously requested their on-demand deposits.  Even the thought of default in people's mind is a self-fulfilling prophecy, since it would cause a chain reaction of other depositors worrying about the security of their deposits, and likewise would request to withdraw their deposits from a potentially untrustworthy bank.  Murray Rothbard does a great job explaining this mechanism in Chapter VIII "Free Banking and the Limits on Bank Credit Inflation" of his book The Mystery of Banking (you can read the entire book for free).  Basically, he argues that the threat of bank runs are a good thing, since they will keep the banks honest (by engaging in 100% reserve holdings).  Here are the relevant paragraphs:


    "The bank run is a marvelously effective weapon because (a) it
is irresistible, since once it gets going it cannot be stopped, and (b)
it serves as a dramatic device for calling everyone’s attention to
the inherent unsoundness and insolvency of fractional reserve
banking. Hence, bank runs feed on one another, and can induce
other bank runs to follow. Bank runs instruct the public in the
essential fraudulence of fractional reserve banking, in its essence
as a giant Ponzi scheme in which a few people can redeem their
deposits only because most depositors do not follow suit."
...
    "Fortunately, the market does provide a superb, day-to-day
grinding type of severe restraint on credit expansion under free
banking. It operates even while confidence in banks by their cus-
tomers is as buoyant as ever. It does not depend, therefore, on a
psychological loss of faith in the banks. This vital restraint is sim-
ply the limited clientele of each bank. In short, the Rothbard Bank
(or the Jones Bank) is constrained, first, by the fear of a bank run
(loss of confidence in the bank by its own customers); but it is
also, and even more effectively, constrained by the very fact that,
in the free market, the clientele of the Rothbard Bank is extremely
limited. The day-to-day constraint on banks under free banking is
the fact that nonclients will, by definition, call upon the bank for
redemption."
...
    "The beauty and power of this restraint on the banks is that it
does not depend on loss of confidence in the banks. Smith, Jones,
and everyone else can go on being blithely ignorant and trusting
of the fractional reserve banking system. And yet the redemption
weapon does its important work. For Jones calls on the Rothbard
Bank for redemption, not because he doesn’t trust the bank or
thinks it is going to fail, but simply because he patronizes another
bank and wants to shift his account to his preferred bank. The
mere existence of bank competition will provide a powerful, con-
tinuing, day-to-day constraint on fractional reserve credit expan-
sion. Free banking, even where fractional reserve banking is legal
and not punished as fraud, will scarcely permit fractional reserve
inflation to exist, much less to flourish and proliferate. Free bank-
ing, far from leading to inflationary chaos, will insure almost as
hard and noninflationary a money as 100 percent reserve banking
itself."


So there you have it.  In a free banking system (which bitcoin is, since there is no central authority), competition amongst bitcoin banks will lend itself towards 100 percent reserve banking.  BitCoin users could conceivably come up with a similar strategy to maintain bank honesty, whereby depositors to a certain bank will deliberately and simultaneously request to withdraw their assets from a particular bank at a randomly chosen (but mutually-agreed) time.  If the bank is honest, it will pass this test.  If it is not honest (i.e. it engaged in fractional reserve banking), it will fail this test.  Or alternatively, a bank could instead not even offer on-demand withdraws, whereby it would maintain a strict policy by having the schedule of payments for loans it lent out to borrowers match the schedule of interest payments to its depositors.

But with bitcoin, unless you wanted to gain interest at real risk of loosing your deposit, there is no need to store your bitcoins at someone else's bank, since your wallet is your own bank:).  Unlike with real species or physically holding fiat currency notes, your bitcoin wallet is not at risk of being stolen by common criminals if you keep it encrypted and backed-up in multiple places online.  All you will have to know is your password, which you can memorize.  But of course you would need a password or pin when accessing a regular bank anyways, so memorizing a password is not an extra burden on you.  So there is no serious need to deposit your money at some bank, if all you desire is to hold your money (without gaining interest).  But since there is not significant inflation of the bitcoin money supply, one need not be concerned with gaining interest simply to keep up with inflation.  So just keep your money in your own wallet:).