The Stellar/Ripple approach to this problem is to replace the single person with multiple people who vote on the ledger. The hope is that while a single person can be easily corrupted, it is much more difficult to corrupt multiple people.
At present, if you pull the code from github.com/stellar/stellard, it will by default install the following "validators". "validators" correspond to pub/priv keys, and under the default config, if 3 of 5 of them "approve" a ledger, then the client considers that ledger valid:
# The latest validators can be obtained from
# https://gostellar.org/stellar.txt
#
[validators]
nakuKsirzNq66iejBfxKxt8Rw4t7kP5ZWSrvWzVKEVSHSkJKYd1 SDF1
n3gVwaSDBtVi4Xd2XBh7rvcwis4uBabu5aNn7WKtEqazJbLHR9n SDF2
nfozTPzMytTKuuHBJrPB9DVxPccKeNVh8vsEomUAwNWuWrNxwDB SDF3
naSBAykwFkmw6JLKqPG7HPFMK1GZqu1rdnwbhtTxyTmVX1b6b8z SDF4
nf3yCmpieMFvAbv5r5jBhGeXheuhc3boAkdmBbpXb6HC7HDkf13 SDF5
# Ditto.
[validation_quorum]
3
Of course the standard Ripple/Stellar response is to say "but you can run your own node and get people to use yours instead!" Which is of course true, except that the consensus mechanism diverges when the network contains nodes with disjoint validator lists and there are hostile actors present.
So for the system to work at all, all nodes much share a common subset of validator nodes in their lists, which in practice ends up being the default set, which are given de facto control the transaction selection.
Which is to say, Ripple/Stellar is not significantly better than the PayPal model.
They are also focusing heavily on getting non technical users. (As evidenced by using facebook as a free funds distribution method.)
And of course those founders are always here on HN ready to downvote any criticism.