Ryan's Journal

New tactics I've noticed in my spam catchers. On the web, I recently started getting trackback intended to entirely confuse learning systems (probably due to the high profile of Akismet). The text is entirely innocuous ("this is a good article") and links to the front page of major sites like Google, MSN, and Yahoo. That's it. There's no pitch, no link to product. Like email spam, it appears we're entering into a phase of combatting learning systems. I ended up deleting them instead of training spam on that. I guess I only consider that spam in the sense that it's unwanted, not that it's offensive or totally unrelated, but it's going to be annoying if I get a flood of meaningless confusion spam.

In my email, I've been getting a lot of the same meaningless, no-sale spam, but I've also started getting a lot of image-based spam, where the textual body of the email is gibberish, and the actual pitch is all in an image (I think; I've never opened those images). This gets past all of my defenses most of the time.

I think their goal is to inspire research and solutions into recognizing image spam. Because then they can go defeat capture systems ("retype the wavy text in this image into this text box") with publicly accessible, readily available code we developed to beat them, and then they can go back to using free mail accounts to send more spam, etc., ad infinitum.

This really is evolving into a tech war.

You must login to leave a comment