Twitter: the next generation of spam
07 August 2009
The moment anybody invents a communications technology, somebody else invents a way to hijack it for advertising. Then people find a way to block those ads, and the spammers make their spam harder to detect. And the battle rages on.
On Twitter, until recently spam has consisted of people setting up a profile, putting an advert in their profile and then following people. That results in an ad view firstly when the you visit to check out who your new follower is. And secondly, when anybody viewing your followers list clicks on them. You can often detect this spam fairly easily - it tends to have few tweets and meaningless usernames (the second of those being much harder to detect automatically).
You can save time validating followers by just leaving it a week before clicking on your new follower notices - Twitter is getting good at banning suspicious accounts based on how many people block them, although if everybody sat back and did nothing, I guess it would fail.
In the last week or so, I've noticed a new type of spam, though, which is harder to detect. It appears to be harvesting random snippets of content from the web or from other tweets, based on a particular search keyword or group of keywords. It seems to be an adaptation of the software that creates spam blogs by lifting content from other blogs.
These new spam Twits can appear to be reasonably valid accounts - it's often difficult to tell the difference between someone with poor language/tweeting skills and a random phrase lifted from a webpage if you're only looking at 140 characters. The thing that first tipped me off to this spam was the bizarre nature of one of the accounts (all about dentistry), rather than the quality of any individual tweet.
It will be interesting to see how Twitter combats this. The strength of Twitter is partly that it enables you to post from any device or program. That makes it particularly vulnerable to spam programs, and it might prove difficult to detect when they're being used. And should they be banned anyway? I can see a valuable role for a program that would, for example, automatically seek out news about Pink Floyd and tweet it. The whole premise of Twitter is that people are allowed to think up new ways of tweeting, and we might find some potentially useful software emerges from spammers' innovation.
(See also Inside the mind of a spammer)
Labels: social networking