Our Algorithmically Stupid Future

Read Time:4 Minute, 6 Second

I am attempting to write a historical novel set in early modern Nuremberg centered around the daughter of a deceased clockmaker who wants to keep her father’s shop and work on early prosthetics (a real thing, by the way) despite the rules against it. And, of course, there is murder, religious, and class conflict, because what’s a story without at least a hint of a guillotine? Think Hangman’s Daughter crossed with Hidden Figures or Radium Girls. As a result, I have been doing a lots of research. One of the tools I have used is Academia.edu. The site has a free and paid tier, and it desperately, desperately, wants me to upgrade. No surprise there, but how it goes about it is one of the more amusingly stupid artifacts of the algorithmic age.

For those not aware, academia.edu is a site that lets you read academic papers and chapters from books. The free tier lets you download certain papers as PDFs and that is pretty much it. Now, I haven’t needed much beyond that as I have had a great deal of success in finding what I need via interlibrary loan (librarians are awesome, by the way. Just though I should mention that.), JSTOR (they let you look at 100 articles per month for free ever since Covid) and, when all else fails, asking professors nicely (historians are also awesome). Academia.edu is not satisfied with that, however. They have money to make.

One of the ways they encourage you to sign up for a paid profile is to allow you to see mentions of yourself in papers. For months, now, I have been getting emails asking if the reference to my first initial last name in such and such paper type could be me. Pay the fee and find out who is referencing you in academic papers! Well, I am pretty damn sure no one is referencing me in academic papers. And if they are, it is likely for something I did that was really stupid, so I am probably happier not knowing. I don’t blame academia.edu for making the attempt. I imagine most of the people using the service are academics of one kind or another, not aspiring historical novelists. But the way it goes about it is somewhat less than optimal.

You might think that, given I have been researching early modern Germany, that it would assume I worked in that field. Nope. It has asked me if I am the person being referenced in papers regarding, among other things, biotechnology, computer science, linguistics, humanities (which might be include history), and the most weirdly specific — Industrial Revolution Papers.

Now, it has been months since I signed up so you might say, “Well, it started with humanities and since you didn’t respond branched out.” Nope. It has used the shotgun approach from the beginning. Right from the moment academia.edu decided I was a candidate for monetization, it sprayed me with every reference to all my academic doppelgangers.

This is not an effective tactic, one would think. It does highlight that algorithms are only as intelligent as we make them. Garbage in, garbage out and all that. But what is more depressing is that the algorithm is probably not that stupid. It likely does work, at least to some degree. Now, the cost of sending emails is really low, so the return does not have to be very high, and the scattershot approach apparently crosses that line. It is perfectly representative of the crappification of much of the internet.

We live in a world where much of what we see in terms of entertainment to news to advertisements is decided by algorithm. And those algorithms are tuned to the bare minimum, to give us not the best for us but just the level of crap that we will tolerate and still hand over our time and money. The same thought process that produces a series of emails that thinks I am both a biotechnician and a historian of the industrial revolution is the same process that shows me the same two dozen shows in my Netflix recommendations and drives irritating news into my social media feed. We are that the mercy of programs that know precisely how much garbage we will put up with before our wallets close and they won’t give us a single thing better.

That is a real loss. Automation has brought us many benefits, but in certain areas it has made our lives worse. Recommendation is one of those areas. We have gone from a world where people think about what would make other people happy if the experienced it to one where machines “think” about how much irritation they can inflict upon people before the money stops flowing. The world is just a little bit rougher, a little bit worse as a result.