Using information bonds as collateral to solve trust problems between strangers
Updated: 27 November 2016 – trust further reduced
I thought I would post an amusing business idea that might be right up Bitcoin’s alley. It’s called Naked Loans and some Chinese students are already using them. Here’s how it works: you snap a naked selfie and offer it up as collateral for a loan along with standard personal information. The lender then approves the loan and you receive the money - fail to repay the money and the lender publishes the nudes. Sound silly? I would have thought so myself until I heard that people were already using them.
Information as collateral is actually an extremely interesting concept that already finds a range of applications. As an example, a high-end online brothel takes compromising information from clients to arrange escorts and it’s been said that certain criminal organizations have used information to minimize the risk of snitching. The point I want to make here is that these kinds of information bonds solve a solid business problem: how to trust strangers over the Internet - and that is quite significant.
If we can solve trust problems with information bonds then we don’t have to rely on large monetary sums to secure agreements, meaning that potentially many new kinds of businesses could benefit from this. But first, you need to establish a legal entity that can accept these kinds of sensitive information on behalf of the contract participants - and that does unfortunately still require trust (though there may be ways to reduce this in the future with blurring, partial information release, AI and so forth.)
One can imagine a legal company that connects lenders and borrows, and handles the process of vetting information bonds for contracts by publishing the bonds on their website if the borrower defaults on the terms I don’t see anything inherently illegal about that if the borrower is over the age of 18 and legally agrees to the contract … I mean, there’s got to at least be similar clauses used in the adult industry already, otherwise production companies could risk getting sued for defamation.
For your added amusement, checkout nakedloans.io. Believe me, I still LOL hard reading this. The marketing almost writes itself.
Removing the embarrassment factor
One of the main problems this scheme suffers from is the fact that the information bonds require a full photo to be submitted to the lender who is then in a position of power against the borrower.
A potential solution to this problem is to cut the photo into three segments:
- Upper-chest - lower face
Segment three includes the lower portion of segment 1 and upper portion of segment 2 such that it can be verified that both segments are of the same body (plus run algorithms to check for photoshop / alterations.) In this way it becomes possible for an information bond to be verified without anyone ever having access to the full photograph.
To build on this idea some more, it may be possible to split the image up into enough pieces so that consensus on the whole can be formed by the feedback of multiple individuals (or AI) working on the pieces. This would allow for a purely decentralized implementation of information bonds to be used as a layer for trust problems in agreements (aside from requiring initial trust in the verification of the persons identity, of course.)
You could even have it so that the verifiers can communicate to describe aspects of the image to see if they line up. I imagine aspects of the image could be transmitted to each verifier (like structural markup and dimensions) to come to a progressive consensus on whether or not the segments are really a whole photo of a single person or a complete forgery.
Double-sided information bonds (inspired by Zimbeck double-sided deposits)
Another scenario is to form agreements directly between individuals who already trust each other in order to enforce some future action. Both parties would meet in real life to exchange information using an application built for the purpose. The software would lock up said information, encrypt it, and upload it to a decentralized storage network. Afterwards, both sides would be required to confirm the contract and the information would become accessible to both sides at the same time.
The game theory dynamics of forming a contract like this are quite similar to the game theory used in nuclear deterrence. Imagine two sides that both have nuclear weapons. Releasing the nukes is devastating to the other side but the moment you launch you risk receiving a counter-strike.
In the case of double-sided information bonds, if a participant takes a negative action against the other side - it might not seem rational to retaliate by release the information bond. But that’s also assuming that you can rely on the other side to stay rational during a conflict - and that I believe may be a credible enough threat to prevent future negative outcomes.
“I call it the Madman Theory, Bob. I want the North Vietnamese to believe I’ve reached the point where I might do anything to stop the war. We’ll just slip the word to them that, "for God’s sake, you know Nixon is obsessed about communism. We can’t restrain him when he’s angry—and he has his hand on the nuclear button” and Ho Chi Minh himself will be in Paris in two days begging for peace.“ - Richard Nixon
To see an interesting historical perspective: Richard Nixon famously employed ”The Madman Theory“ during the Vietnam War to try convince his opponent he was irrational and temperamental. The theory was that if he were to remain rational there would be no way he would do something as insane as nuke an entire country … but if he wasn’t rational …
Other information bonds
The example given at the start is to use nude photographs as the information bond but many other kinds of information will also suffice. Other types of information include:
- Racist, homophobic, or sexist statements
- Negative remarks about friends and family
- Proof of infidelity
- Statements of criminality, fraud, and mistrust
- Damaging your credibility: mental illness, lack of qualifications, etc
Some other use-cases for information bonds
- Decentralized exchange can be done with information bonds - this would allow Bit Square to operate without requiring a monetary deposit to purchase Bitcoins - which is impressive. Trade Bitcoins and every kind of asset in a peer-to-peer marketplace without trusting a central exchange with money (information bonds become the new money, but hey – maybe that’s preferable to losing your life savings? Who knows.)
- Decentralized marketplaces like Open Bazaar can be done with information bonds.
- All current contracts supported by Bithalo would work with information bonds.
- Organized crime becomes much easier to do (for better or worse, these things usually swing both ways.)
- Prostitution marketplaces become possible if a reliable escrow agent can establish a reputation (FYI: this is already happening.)
- All current agreements that require money to enter into escrow can be replaced with information bonds.
- A more novel use-case is to establish a criminal reputation system using SSL for proof-of-crime. Researchers recently described a cryptocurrency whose proof-of-work was based on DoSing websites running SSL. TLS-notary also allows for proving the contents of a browser so these could be applied to produce cryptographic proof-of-crime that can be used as information bonds.
Compromising information can be used as a bond to secure agreements between untrusted strangers in the place of monetary collateral. This idea can be applied to a number of uses-cases to improve usability, reduce costs, and prevent the risk of default in agreements. However, because of the risk for blackmail a third-party must be used to enforce these agreements. The scheme can be further improved by splitting up information between oracles such that the information can still be verified without requiring that the oracles have access to the full information bond.
- Future research directions may include producing cryptographic proofs that prove that all of an images sections match the requirements for an unaltered image of the same person.
- How can a person’s identity be established with minimal trust?
- Can the concept of information splitting work with other kinds of information?
- Can information splitting be cheated (and how can this be detected?)
- Can this be done purely peer-to-peer using a reputation system without the risk of collusion?
- Can information blurring work to reduce trust without reducing verifiability?
- Can provable deletion be accomplished with pub-key crypto and trusted computing?
- Is it possible to calculate individual anxiety of an information bond being release relative to the financial gain using prior psychological testing (this is a big problem.)
- Can legal contracts be used with the above statements to ensure that a person will care if an information bond is released (or didn’t lie to terms.)