Not that stupid

commentary I'm afraid to admit it, but it looks like spammers are winning the spam war. But who actually falls for all those too-good-to-be-true messages?



commentary I'm afraid to admit it, but it looks like spammers are winning the spam war. But who actually falls for all those too-good-to-be-true messages?

Let's face it: someone out there thinks we are pretty damn stupid. But how dumb are we?

Social researcher Hugh Mackay wrote in the Sydney Morning Herald  a few weeks back that in this age where trust of politicians and (of course) journalists is at an all-time low, by contrast advertisers are viewed as an almost trustworthy source of information. His reasoning goes that at least with advertisers you know where you stand--they're trying to sell you something.

Mackay obviously doesn't have e-mail, or he'd have a thing or two to say about advertising being on the level. The level of spam and hoax e-mail is approaching the seriously unworkable.

In previous columns I have put my faith in those clever people with glasses who developed spam filters based on heuristic analysis (also called Bayesian filtering). Now it seems the spammers have hired some clever people with glasses of their own. First of all, there's the obvious trick of misspelling words, as in "v1@gra" or "b00b1es", which most well-designed spam filters can catch. But then, has anyone noticed the large number of spam messages that include at the end enormous lists of meaningless words like "drippy ancestor committed wrapup redactor groundwork" or "abstain sour boggy edwin illogic illicit forbore preferential adulate"? (With my luck, the spammer I copied those word lists from will try to sue me for plagiarism, so just in case, I'll give them credit--it's that company selling the CD of secrets the government doesn't want you to see.)

As far as I can work out, this is a very clever way of throwing out those heuristic tests. To you or me, it looks like complete gibberish, but to a computer doing statistical analysis of word frequencies it looks like a whole bunch of legitimate words. Any naughty words like "free holiday" or "thicker and fuller" will be statistically insignificant in comparison to the huge number of "proper" words.

Is there a way to teach computers to recognise words in proper sentences, rather than long strings of words with no meaning? Probably, but for now it looks like the clever spammers are one up on the clever spam preventers. It's depressing.

The thing I don't get is, these messages with enormous strings of gibberish are so obviously spam I can't imagine anyone thinking they are legitimate e-mail. So while the spammers have defeated the spam filters, nobody is actually going to take their messages seriously. Right? Or have spammers given up on selling us things and now their only motivation is a bloodyminded obsession with getting into our inboxes, annoying the crap out of everyone, and making e-mail more of a burden than a business tool?

It's just that with the exponential growth of hoaxes and spam out there, you have to wonder--someone out there must fall for them, surely, or there would be no point.Everyone's heard that story about overhearing some guys in a pub who fell for a Nigerian or 419 scam. Everyone has that annoying friend who forwards hoax e-mails, adding "I don't usually forward these, but . . ." Every month or two, we at ZDNet Australia get an inquiry from someone who's received that hoax e-mail--the one with the cute picture of the baby wrapped in a ribbon "from God", that says AOL and ZDNET will give $1 for every person who forwards it--asking if it's for real. But who are these people, and is there any hope for them?

Getting back to Hugh Mackay, just as we have become more cynical about political and media messages and better able to deconstruct them, surely some day collectively we will get wise to spam and hoaxes, and then there will be no point sending them. Wouldn't that be wonderful? Or was PT Barnum right about suckers?

Subscribe now to Australian Technology & Business magazine.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All