An interesting speculation from Charlie Stross (who needs to be on silk)
- I've seen versions of this idea before, but this is probably the best
articulated.

excerpted from
http://www.antipope.org/charlie/blog-static/2010/12/its-made-out-of-meat.html

Spam is everywhere.

About 92-95% of all email traffic is spam. Every new communications
medium that opens up on the internet succumbs rapidly to spam, unless it
is designed with such heavy filtering in place that it's almost
impossible to send a message to someone else without prior approval. But
new communications media don't get adopted unless they're useful — and
one of the key uses of a communications medium is to allow strangers
with useful information to get in touch. Spam, almost by definition,
isn't useful: but it tries to masquerade as meaningful communication.

In the bad old days of email, just about everything anybody sent would
eventually get delivered to a mailbox, if it was correctly addressed.
When the "anybody" using the internet expanded sufficiently to include
unscrupulous advertisers and scam artists, the utility of email began to
drop. The solution that eventually turned up was the widespread adoption
of filters — software that attempts to determine whether an inbound
message is unsolicited rubbish, or something potentially of interest to
a human recipient.

There are a vast number of ways of filtering. One of the most effective
is to look for patterns in the mail stream; an identical message sent to
a million people is almost certainly spam unless it emanates from a
well-known mailing list system. Unique messages are less likely to be
spam. So looking for huge deluges of identikit mail worked for a while —
until the spammers took to appending random snippets of text to each
individual message, to make them look different.

Another filtering technique is to look at the word or letter frequency
of the message; purely on a statistical level, spam doesn't look like
part of a conversation (unless your correspondents regularly interrupt
the flow of discourse to shout BUY CHEAP DESIGNER HAND-BAGS or similar).
But again: spam is big business — it's a very effective form of mass
advertising — and the spammers are ingenious.

As filters get more sophisticated, the spammers are abandoning old-style
broadcast advertisements and are moving to much more tightly targeted
ads, addressing the recipient by name and attempting to pitch
selectively. The most tightly targeted spam is created for spear
phishing attacks (in which specific personal information is used to
target selected individuals — usually for identity theft or corporate
espionage). Today, this is labour intensive: but it's a fair bet that as
more of us place more information about ourselves online, spear phishing
techniques will gradually become automated, and targeted junk internet
advertising will rise to levels of sophistication we can barely guess
at. There's lots of money in spam (these days it's a branch of organized
crime), and where there's money, talent can be hired.

We are currently in the early days of an arms race, between the spammers
and the authors of spam filters. The spammers are writing software to
generate personalized, individualized wrappers for their advertising
payloads that masquerade as legitimate communications. The spam cops are
writing filters that automate the process of distinguishing a genuinely
interesting human communication from the random effusions of a 'bot. And
with each iteration, the spam gets more subtly targeted, and the spam
filters get better at distinguishing human beings from software, in a
bizarre parody of the imitation game popularized by Alan Turing (in
which a human being tries to distinguish between another human being and
a piece conversational software via textual communication) — an early ad
hoc attempt to invent a pragmatic test for artificial intelligence.

We have one faction that is attempting to write software that can
generate messages that can pass a Turing test, and another faction that
is attempting to write software that can administer an ad-hoc Turing
test. Each faction has a strong incentive to beat the other. This is the
classic pattern of an evolutionary predator/prey arms race: and so I
deduce that if symbol-handling, linguistic artificial intelligence is
possible at all, we are on course for a very odd destination indeed —
the Spamularity, in which those curious lumps of communicating meat give
rise to a meta-sphere of discourse dominated by parasitic viral payloads
pretending to be meat ...


-- 
((Udhay Shankar N)) ((udhay @ pobox.com)) ((www.digeratus.com))

Reply via email to