A challenge for Mollom
When I was checking the Akismet moderation queue of our (old) photo blog, I noticed that spammers are doing their best to outsmart automated spam blockers such as Mollom.
Here's a screen shot of the spam message I bumped into:
To be honest, it fooled me when I saw it passing by, and I almost approved it as a genuine message.
The trick they're trying to pull here is to simply copy a previous comment, and add a hyperlink somewhere in it. That way, the user would be tricked into following a link he/she believes might be interesting.
I don't know if this is done is any automated way, but I'm probably not the first to notice this kind of spam. I wonder if a service like Mollom would be able to recognize it. Time will tell...
Comments
Any updates on how it's going
Any updates on how it's going now?
Normally, Mollom should be
Normally, Mollom should be able to catch this rubbish. I have had very few false negatives over the last year, since Mollom went into beta. The classification algorithm is pretty good, and if it's unsure, a captcha will be generated. The only unknown in the system is how well machines are able to outsmart captcha's. But as long as it does not pay off, I doubt the spammers will invest a lot of effort into breaking them.
Post new comment