Note that the security level of this by design is no more than the proof-of-work cost of N blocks, if you pay back the loan N blocks before the deadline. Proof:

Note that the security level of this by design is no more than the proof-of-work cost of N blocks, if you pay back the loan N blocks before the deadline. Proof: If the required key is X bitcoin blocks with no transaction to lender, and you pay it back after X - N blocks, the lender need only mine N blocks, starting from the Xth, to obtain the key (X blocks). They can do this at any time. Therefore, to obtain "acceptable" security, one should probably pay back the loan long enough before the deadline that the economic cost of mining N wasted blocks outstrips the value of the blackmail material itself. (Not considered: the economic calculation probably changes if one entity is giving out multiple such loans, and can therefore mine "fake" witnesses for multiple encrypted materials at once?)