I was amused to receive an automatic warning from Google recently. This email kindly described their automatic blog spam detection program, and said that my condo blog had been flagged as a possible spam blog. The email said, "If you're a human, just click this link and follow the instructions. Someone will review this blog and take it off the list."
I chuckled and complied. It took them about 48 hours to approve it, during which time C to the Fifth was still accessible, but I was not allowed to add or edit any posts.
I'm still unsure exactly how I feel about their quote-unquote algorithm. On the one hand, I'll cheerfully grant that californiacondo.blogspot.com is a URL that just screams spam. I wouldn't be shocked if Google was flagging every blog with the word "condo" in the title - there's a huge number of web pages out there of very questionable value, all hawking real estate of some kind.
On the other hand, I can't imagine any decent algorithm that would process the actual content of my blog and conclude that it's spam. For one thing, there are almost no links there (at least, there weren't until my recent post-spam-block post about local mortgage lenders), and links are the main reason spam blogs exist in the first place. And while recent spam generators have gotten pretty good at mimicking human writing (at least in terms of words per sentence, paragraph structure, punctuation, and word variety), they still aren't as good as real humans... I would have hoped that Google could tell the difference between my writing and that of a bot.
Anyways, just another interesting chapter in my ever-evolving relationship with Google. I'm glad that they're tackling this problem, just skeptical about the quality of tools they're using to address it.
No comments:
Post a Comment