An Algorithm is Watching You

 Annamaria on Monday

This past week I googled this question: "Can a crocodile smell blood?"

As you may have surmised from seeing my recent posts about the air war in East Africa in 1915, I am working on the fifth book in my Africa series, to be called The Sharpest Tooth.  Yes!  My story involves, in addition to the regular series characters, a young sailor aboard one of the combat vessels.  When I asked the crocodile question, I already knew that a pilot who crash-landed in the Rufiji River in 1915 and disappeared was assumed to have been eaten by a croc.  I wanted to know if such an event would  bring more crocs near my young hero floating on a river monitor barely a two feet above the water.  Google gave me this answer:

Then I began to wonder what the spying bots would make of the kinds of things that I and my fellow MIE authors type into their search engines.

I had been aware for some time that the things crime writers talk about can be misleading to eavesdroppers.  I had, for instance, participated in such a conversation.  Years ago, when I was the president of the NY Chapter of Mystery Writers of America, I had to co-sign the chapter's checks with treasurer.  She worked in Times Square, and I researched at the New York Public Library a couple of blocks away from her office. Once a month, we met for a coffee so we could sign together.  And it gave us a delightful opportunity to chat.  On one such day, she was telling me about a problem she was having with a plot:

She: Do you know an expert in poisons?

Me: I have a book about them at home. I'll be glad to lend it to you.

She: Great.  I want it to look like suicide, so I need to find a poison that is not at all easy to detect.

With that, the people sitting near us moved to another table. 

When she and I laughed, it probably only cemented their suspicions.

But what would an eavesdropping algorithm make of such an exchange.  We all know that bots belonging to search engines are spying on our every move on the Internet.  

(It occurred to me as I typed the above sentence, that the letter "I" in AI does not refer to intelligence, as in the capacity of big-brained humans to solve problems, make decisions, etc. etc.  It is really much more like the "I" in CIA and MI5.)

I began fretting over this issue years ago when I was researching Idol of Mombasa.  Each book in my Africa series has a plot that involves one of the Ten Commandments, and it also has another sin, not forbidden by Biblical law, but one that I have the audacity to think should have a Commandment.  In Idol of Mombasa that other sin is slavery, and a lot of the story's tension stems from the conflict between British law and Sharia law.  By 1912, the Brits had outlawed slavery, but Sharia, the law embraced by the Muslim population, did not.  And Mombasa, newly taken over by the Brits, had been an Arabic city for a couple of millennia. To write that book I needed to learn a lot about Islam in East Africa

I was beginning that research in 2014, when the US was deeply involved in a war in Afghanistan, and there was a lot on the news about home-grown young men being recruited for terrorism. And there I was googling question after question after question about what Sharia law said about x or Y.

Do you think it wrong of me to imagine that there might have been a bot watching?

More to the point of MIE, what do you suppose the spying bots think of the questions you ask?