Thesaurus Spam 2: The Comment Years
“Thesaurus spam” tries to avoid automated unsolicited-commercial-message detection by automatically replacing words in the spam text with “synonyms”. I put scare-quotes are around “synonyms” because thesaurus spam often fails to pick anything even close to a true synonym. So “we will fight them on the beaches” could, for instance, become “ourselves will affray them on the littoral”.
I hardly receive any thesaurus-spam via e-mail any more (largely because of upstream filtering; it’s probably still quite popular), but I do still see it. Most recently, in comments on this blog.
What happens is, a spammer comes along and creates a commenting account with a “Website” link to whatever site they want to spamvertise. Today, this was a commenter called “batterysea”, linking to www.uk-power-battery.co.uk. (All evidence of this commenter has now been erased, of course.)
Then the commenter goes into robospam mode. Instead of posting the usual robospam comments that say something like “Louis Vuitton Prada best replica fakes Rolex Viagra” et cetera et cetera, with links to a Web site from pretty much every word, they create an innocuous, linkless, plain-text comment. At a glance, the new spam-comment kind of looks as if it belongs on the page. That’s because it does kind of belong there, on account of being a copy of an earlier comment on the same page, but with the Thesaurus-O-Matic run over it to make the copying less obvious (and difficult, if not impossible, to auto-detect).
I’ve plucked a few of these ticks off the blog before, but this one this one managed to splatter a few more comments around before I stopped him, so I paid more attention. I presume these spammers try to strike a balance between getting a commercially useful amount of spam transmitted, without obviously producing tons of new comments that even a dozy admin is likely to notice. In the “batterysea” case, there were nine comments, posted at one-minute intervals on my nine most recent posts.
On this post, for instance, there’s a legitimate comment from Anne that says
Clearly I am culturally deprived - I don’t read magazines, I don’t watch TV, and I surf the web with adblock. So where would I see these ads?
Maybe a better question is, do these ads actually sell products? I mean, if I’m trying to decide on which fan to buy for my PC, is seeing an ad in a magazine actually going to affect my decision, whether the ad has giant robots or sober statistics?
And then, at the end of the page, along came the spammer to say
Clearly I am culturally beggared - I don’t apprehend magazines, I don’t watch TV, and I cream the web with adblock. So area would I see these ads?
Maybe a more good catechism is, do these ads absolutely advertise products? I mean, if I’m aggravating to adjudge on which fan to shop for for my PC, is seeing an ad in a annual absolutely activity to affect my decision, whether the ad has behemothic robots or abstaining statistics?
On this post, the spammer lifted just the second paragraph of my own comment, which started out
It’s possible that such a scheme would actually be legit, but it’s probable that it would not, because people sending money would have the implicit assumption that they were going to get something in return, even if it was as unlikely to be valuable as a lottery ticket.
That part became
It’s accessible that such a arrangement would absolutely be legit, but it’s apparent that it would not, because bodies sending money would accept the absolute acceptance that they were activity to get article in return, alike if it was as absurd to be admired as a action ticket.
…in the spam-comment.
When the robospammer can’t find any words to thesaurusise, it ends up just duplicating an existing comment. For instance, Fallingwater’s comment on this post:
The Asus EeePC 1005HA is, I think, the device that loses its rubber feet fastest than anything else that has been produced.
My solution: melt glue. Four puddles where the feet used to be have made my EeePC stick to surfaces again. Less than when it had the rubber feet, but a hell of a lot better than naked plastic.
…was duplicated word-for-word by the spammer.
This is a really feeble kind of spamming. All commenter Web-site links on this blog, and pretty much every other blog, are nofollowed, as are links in the comments themselves. So you don’t get search-engine prominence from this technique, and you don’t even get any traffic to speak of, unless human readers click on your commenter-name. I presume this happens even less often than people clicking on the links in the “Dolce Gabbana Dior bags Gucci handbags Chanel Hermes…” sorts of comments.
I think the only way to make comments that really look as if a human posted them would be by creating a spambot with something resembling real, “strong“, AI, like the burgeoning network-creatures in Maelstrom, the second of Peter Watts‘ excellent “Rifters” series (all three books of which are downloadable for free!).
In the meantime, we get aphasic thesaurus-robots, all that can be said for which is that they’re more successful than the robots that make hundreds, and hundreds, and hundreds, of accounts called things like “aFZflRhBzRsYq <asdfwerj5@gmail.com>”, but never manage to post a single actual comment.
