Spam Bots and CAPTCHAs

Last time I told you about the automatic spam killer called Askimet. Although this nifty little plug-in seems like the much-awaited answer to the ever-growing problem of spam, it really isn’t.

Here’s the dark side of Askimet

I believe that determining spam is one of those situations where there isn’t an acceptable substitute for human intervention. There’s no doubt Askimet saves time by processing comments, trackbacks or pingbacks blazingly fast. There’s also no doubt that it sometimes makes the right calls.

But no matter how much its knowledge base expands, Askimet won’t ever be able to distinguish spam from legitimate comment 100% of the time. Heck, it might not even get it right 50% of the time.

The creators of Askimet know this which is why anything it thinks is spam isn’t immediately deleted. Instead you’re allowed the opportunity to “check over”? the results and make any adjustment necessary. They don’t want to be responsible for deleting something YOU might otherwise feel is important, perhaps even crucial to the success of your business.

If you still have to look over Askimet’s shoulders maybe you should ask why even bother with it at all?

Nowhere is this problem more apparent than in posting criticism. Usually if you’re going to post a comment you’re required to sign up for the blog. As humans we’re typically less inclined to post comments, especially negative ones, if we’re required to attach our names. Most will choose not to comment. Those who do post criticism risk having a webmaster or Askimet flag it as spam anyway. Either way, the comment doesn’t get read. Where’s the good in that?

Here’s something else to consider: If Askimet can learn over time to distinguish good from bad, can it be only a matter of time before spammers learn how to “fool”? Askimet’s logic and defeat its very purpose? Spammers are a determined group and when there’s a will there’s a way.

Now that’s something to think about…