After my last entry about spambots, I noticed quite a few of them ‘Liking’ the Agnoistrology.com Facebook page. One of the characteristics they all shared was an Egyptian origin.
Now, many were obviously not human; they listed their hobbies and favorite shows and movies in generalities like ‘Hiking,’ ‘Action Movies,’ and ‘Reality Shows,’ without any specific examples. In addition, many of the FB pages they ‘Liked’ were huge, and I mean international huge, corporate pages designed only to advertise a product line. Did I mention that they also invariably listed the Bible or the Koran (or both!) and no others as a favorite book? It became apparent early on that these profiles of young adults wanting to meet both males and females while pursuing a higher education were artificial personalities constructed to boost the fan count of certain pages. I’m guessing they became fans of my page randomly in order to appear to have unique interests.
Unfortunately, I soon found myself automatically assuming that any new Egyptian fan of Agnoistrology.com was a spambot. I was profiling FB profiles!
This got me thinking about Turing tests.
The only real criteria I was using to judge the humanity of these profile owners was the unlikelihood of someone displaying all of the characteristics of a herd mentality also expressing a fondness for my site. My conscience played devil’s advocate: One of these kids could have a relative working for that soft drink company and just think this pop singer is attractive…and have a literary interest in that religious book…and know someone on that reality show…and work part-time for this fast food chain. Who am I to ban them from Agnoistrology? Let them broaden their horizons!
I was unwilling to ‘Friend’ any of them, however, without a reliable, automated Turing test. Ultimately, I’m confident I would have definitively and accurately categorized them with a minimum exchange of messages but what would have really been handy is a CAPTCHA of some sort.
But how long will CAPTCHAs be viable? Frankly, I’m surprised they still work as well as they do. Speech recognition has gotten pretty good and a Japanese research group is already developing chemical sensors that utilize a synthesis of bio- and computer technology to react to certain smells. It seems to me that a paradigm shift is looming that will leave Turing tests behind and I probably should have seen it coming the first time I was fooled into holding a conversation with someone’s answering machine.
