how cranks cheat algorithms and silence their critics
While recommendation and ranking algorithms may be a necessity to navigate our digital world, they’re not without their faults. In fact, you can argue that many of them are broken if we look at social media and virtual bookstores. Just consider what Amazon did when it assembled a list of books on epidemiology, the science of diseases and how they spread. An article in Wired starts with an assessment that can only be described as scathing, noting how much blatant, obvious anti-vax propaganda deceptively masquerades as actual medical research, and just how easily one of the world’s largest marketplaces falls for the ruse.
One has a confident-looking doctor on the cover, but the author doesn’t have an MD — a quick Google search reveals that he’s a medical journalist with the “ThinkTwice Global Vaccine Institute.” Scrolling through a simple keyword search for “vaccine” in Amazon’s top-level Books section reveals anti-vax literature prominently marked as “#1 Best Seller” in categories ranging from Emergency Pediatrics to History of Medicine to Chemistry.
The first pro-vaccine book appears 12th in the list. Bluntly named “Vaccines Did Not Cause Rachel’s Autism,” it’s the only pro-vaccine book on the first page of search results. Its author, the pediatrician Peter Hotez, a professor in the Departments of Pediatrics and Molecular Virology & Microbiology at the Baylor College of Medicine , has tweeted numerous times about the amount of abuse and Amazon review brigading that he’s had to fight since it was released.
Things don’t get better from there as book-length ads for juice cleanses and alt med snake oil co-mingle with serious books about cancer, and conspiracy potboilers jockey for position over legitimate historic and current events tomes. Certainly there’s a market for both dangerous woo and paranoid screeds, as the continued existence of media ventures like InfoWars, Goop, and The Dr. Oz Show will attest, but it can’t be so large that it buries almost every important rubric in a torrent of nonsense, can it? What’s going on? Is there no one in charge? Aren’t there any real standards for what can and can’t be sold?
where did the algorithms go wrong?
Well, the answers to all these questions are far from comforting. It’s entirely possible that few Amazon employees know exactly what heads what section of a book category because their job is to manage the logistics for selling and delivering more than 3 billion different items to a global customer base, and the algorithms who rank anti-vaccine tirades over actual science are meant to figure it out for them. Unfortunately, the algorithms are built to measure engagement and the keywords used by sellers to build their rankings. It doesn’t matter that a book is correct or even what it advertises itself to be. All that matters is that it’s popular and the people who buy it like it and leave positive reviews.
You can think of it as the tech version of “all press is good press” and “doesn’t matter what you say as long as you’re talking about me” schools of thought. Engagement is good. The topic and the reason behind a spike in engagement? Irrelevant unless the media and politicians make a big deal about it. And furthermore, as far as many tech companies see it, there’s nothing to fix because nothing is broken. People are using the sites, buying the products, and reporting what they don’t like. Where’s the problem? Surely, if too many customers hate problematic content, they’ll report and block it. If there’s an issue with the products, they’ll avoid them and complain, and the algorithms will catch this disapproval, ranking and guaranteeing accordingly.
fake it till you make it, social media style
Unfortunately, those who want to abuse the algorithms have figure out exactly how to game them. Use the right keywords to solicit clicks and interest, sending what you’re trying to sell or advertise trending or viral. Get a lot of like-minded fans and bots to provide tons of positive and consistent engagement. Flood the feeds with an ocean of content. Brigade your competition with as many negative reviews and reports as possible. In short, they use everything an algorithm could use to sort content and products in bad faith. This allows bad actors to effectively scream over good, important content, shove it out of the way, and bask in artificial popularity while tech platforms shrug and say “well, guess that’s what the people want.”
But that’s the problem. It’s not really what the people want. It’s what highly motivated cranks and grifters want, and they’ve responded to a dearth of gatekeepers meant to empower those who were too often arbitrarily ignored, if not shunned, by gaming the computers and distorting social conversations and popular rankings. If the tech companies or activists try to step in, the cranks and grifters scream bloody murder about their hard-rigged freedom of speech being infringed on by those they successfully shouted out and would like to keep quiet through brigading, torrents of reports of supposedly inappropriate content, and special pleading with media influencers who are convinced that social media and rankings are reflective of real world trends.
If we’re going to keep living in a digital world where we rely on algorithms to curate what we see and understand the public sentiment, we’re going to need better, more sophisticated, and harder to deceive code, able to understand when users and bots are trying to abuse them to game rankings and dominate the broader discourse. Failure to do so means seeing social media and online marketplaces go the way of far too many comment sections which were abandoned to trolls trying to outdo each other and malicious scripts.