Aggravating Algorithmic Assumptions

I can still recall the first time I felt annoyed when a social media conglomerate thought it knew my brain better than I did. It was back in the early days of Twitter, when I first signed up and chose to follow a dozen or so friends, public officials, and news outlets. Since I was myself a public official of sorts (president of a public university), I thought it wise to stay in touch with these acquaintances as well as voices from across the political spectrum, especially those whose decisions might impact my university.

Then, one day, it happened. I began to notice that the items in my Twitter feed were no longer arriving chronologically. Instead, I learned that Twitter was using some sort of algorithm to determine which tweets I “needed” to see first. That algorithm, I now know (but didn’t at the time), was based on a set of variables that included (a) how much money the tweeter had paid to Twitter for the privilege of placing said tweet in front of my nose, (b) how many times I had “liked” a particular sender’s previous tweets, thereby leading Twitter to the (erroneous) assumption that I would therefore always want to see that person’s tweets ahead of all others, and (c) the extent to which I had “engaged” with that tweeter, as in taking the time to write a comment or instant message.

What happened, of course, is that I no longer saw the tweets from many friends, officials, and news outlets that remained of great interest to me, but whose tweets had never created within me an urge to press the “like” button or otherwise “engage” via the typing of a comment or instant message. Thus, an important purpose in my using Twitter—namely, to remain in touch with a wide spectrum of disparate voices—was frustrated, not by any decision on my part, but because of assumptions made by Twitter’s algorithm generators. They assumed then, as do Facebook, LinkedIn, and a host of other establishments, that if I choose a to follow a particular path for a few days in a row, I must therefore want to create an ever-deepening rut, from which I never want to extricate myself.

Spirited Reasoners are painfully aware of the fact that the example I referred to above is merely the tip of the iceberg. Almost every social media outlet and online retailer draws conclusions about the inner workings of our brains based on what websites we choose to visit, what purchases we make, and what places the GPS in our cell phone tells them we have been.

Viewed in their most positive light, these assumptions can be comical. Consider the following personal examples:

  • A few years ago, my wife bought me a Kindle. Since she was the one who bought it, Amazon (Kindle’s parent company) has always assumed that the choice of books I read should be dictated by an algorithm which is based, in part, on a woman’s online and retail preferences rather than that of the true user (me). Suffice it to say that Amazon frequently treats me to a selection of new Kindle books based on what a woman of her retail proclivities would enjoy. I sometimes wonder how the folks in charge of tweaking Amazon’s algorithm handle my actual choice of books.
  • When I am not writing this blog, I am writing books—currently works of fiction. As with most fiction writers, I typically go online to search for factual tidbits to spice up various scenes. These items bear exactly zero resemblance to anything I would ever actually want to buy. In fact, I tend to search online for them for the very reason that I have never shopped for them myself, nor would I ever want to. I enjoy the fantasy that my online searches are keeping the algorithm folks at Google and Bing busy at nights, trying to figure out my purchasing patterns.
  • Ever since property values began to soar a year or so ago, I’ve been curious about valuations of homes in my own neighborhood and those of friends and family members across the nation. After visiting websites such as Zillow and Trulia to check out listings purely for reasons of curiosity, I began receiving emails, texts, and snail letters from real estate agents in those parts of the country, offering to help me complete my relocation. Over time, I’ve noticed that if I want to receive a half-dozen or so contacts like this, all I need to do is go online and check out the price of a property somewhere else.

On the other hand, when viewed in their most negative light, these types of algorithms can be downright scary. How many nefarious individuals and organizations are keeping tabs on me via my social media and online activities? How many of these might involve foreign or domestic government agencies? How many are using flawed algorithms like the ones described above?

I wish I had an answer to all this. Perhaps its time for Spirited Reasoners to scale back on the use of companies who operate this way. If only we knew how to do it.