Cryptography Reference
In-Depth Information
from the protocol so as to influence the outcome of the survey, perhaps
for political reasons. Equally problematic, however, is getting honest
respondents to perform the protocol correctly: randomized response meth-
odologies can become rapidly cumbersome, requiring respondents “to
perform complex randomization tasks for which interviewers must provide
seemingly 'mindboggling' explanations; the result can be respondent
suspicion, confusion, and uncertainty as to the level of disclosure that a
truthful answer entails.”
57
Moran and Naor's “human-centric protocols” thus bring additional
benefits to the table. Because the security properties of scratch-cards or
sealed envelopes are well understood, respondents are more likely to
believe that the protocols perform as advertised than if they were imple-
mented using computers. As they note, “the computers and operating
system actually implementing the protocol may not be trusted (even
though the protocol itself is). . . . Even for an expert, it is very difficult to
verify that a computer implementation of a complex protocol is correct.”
58
Basing cryptographic primitives on familiar technologies with a long social
history may thus provide a way to develop collective trust in modern
cryptographic protocols while maintaining the rigorous mathematical
analysis characteristic of the provable security framework.
Cognition
Developing such collective trust may require not only transparency of
technical means but also the ability for users to understand how the pro-
tocols themselves perform their often mysterious feats. In fact, the security
proofs of Moran and Naor's human-centric protocols depend on the ability
of honest parties to follow instructions correctly, something they recognize
is difficult to either guarantee or model mathematically. Perhaps, then, the
development of something akin to
cryptographic literacy
is just as necessary
to information security as that of impregnable mathematical protocols.
Researchers tackling this issue have tended to adopt one of two distinct
approaches.
59
The first one frames the issue in terms of “public understand-
ing of science”; that is, the cryptographic literacy problem lies in getting
the general public to understand the capabilities of modern cryptographic
protocols for ensuring security and privacy over networks. Cryptographic
literacy allows the public to “cultivate a higher level of trust for systems
