[ExI] cool, an actual conspiracy!

[Anders]

I recently looked a bit at conspiracies (as a model for information 
retention). There are a fair number that have been documented, like the 
Glomar Explorer. The best data set was however the more boring category 
of business cartels: fairly large sums of money involved, the median 
time from formation to discovery is 5 years (but one persisted for 95 
years). (http://ageconsearch.umn.edu/bitstream/28650/1/wp060011a.pdf)

Generally, conspiracy stability looks bad if it requires large groups 
aware of the conspiracy information. If the probability of leaking per 
year is p, then the probability of a leak per year is 1-(1-p)^N where N 
is the number of members. This is why the broad conspiracies conspiracy 
theorists like are unlikely (see the flawed but fun 
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0147905 
). Compartmentalization might work: I assume for example that building 
the Glomar Explorer could be done with practically nobody involved aware 
of the actual intended use, and then it was just a matter of having the 
crew keep a secret.

So a "conspiracy" of a search engine not finding stuff about someone 
could presumably be done well, since there might just be a handful of 
people messing with the database. The problem is that as soon as it is 
noticed, the force of scrutiny becomes terrible (Facebook and Twitter 
have encountered this).

An interesting look at the problem of keeping stuff secret, from 
somebody in the signals intelligence world, is this:
https://static.newamerica.org/attachments/4425-the-declining-half-life-of-secrets/Swire_DecliningHalf-LifeOfSecrets.f8ba7c96a6c049108dfa85b5f79024d8.pdf

-- 
Dr Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University