Garbled Autonomous Bots Leveraging Ethereum

Garbled Autonomous Bots Leveraging Ethereum
Explainer by Michael P Frank

What we’ve done in this project is to demonstrate how one can achieve absolutely *perfect* functional and data privacy (subject to certain caveats) in programs running on a public blockchain (i.e., smart contracts).

What do I mean by this?  Simply that, if our techniques our used, then (under certain key assumptions) *nobody* who is not authorized can infer *anything* whatsoever about either the data being processed, or the computation being performed on that data.

The technique thus provides a means whereby an automaton may “hide in plain sight” on a blockchain, and it completely destroys the idea that it ought to always be possible, even in principle, to reverse engineer smart contracts to see what they are actually doing.  Indeed, we can rigorously *prove* that, under our basic assumptions, to reverse engineer our machines is is completely impossible.

How is such a capability useful?  Well, if you want to create a bot that is entirely as accessible, reliable, and survivable as the blockchain that it’s running on, and that is either serving some secret purpose on behalf of your organization, or that maybe just needs to be obscured to protect valuable information and methods from being stolen or analyzed by adversaries, this is one way to do it.

In our scheme, even the *users* of the automaton (the entities that feed it inputs and/or read its outputs) can be prevented from discerning anything about its detailed function.

This may all sound impossible, but it turns out to be rather simple to implement.  Naturally, there are some overheads imposed by the encryption methods used, but they are not so large as to preclude some interesting applications from nevertheless being feasible (such as the fanciful example that I posted earlier of auctioning off one’s nation-state to the highest bidder).