KHAP: Using Keyed Hard AI Problems to Secure Human Interfaces

There is often a need for users to securely interact with a remote computing system, ensuring the integrity and authenticity of transmitted messages. Typical solutions assume that a local trusted computing platform is available to perform cryptographic operations, but this is often not the case. We introduce KHAP, a protocol for using hard artificial intelligence problems to provide message authentication checks centered around a human verifier. We also formally introduce the notion of a keyed hard AI problem, which is one that uses an authentication key to prove the source and integrity of a message. We give examples of some keyed hard AI problems, as well as examples of KHAP’s applicability to the specific problem domains of Internet voting and the use of smartcards for digital signatures.