A few years ago, I wrote about how, when planning my wedding, I signaled to the Pinterest app that I was interested in hairstyles and table decorations, and suddenly I was inundated with suggestions for more such events. Everything was fine until – oops – I canceled the wedding and it seemed like the Pinterest pins would haunt me for the rest of my days. Pinterest wasn’t the only culprit. All social media wanted to recommend things that were no longer relevant, and the stench of this musty content buffet lingered long after the event ended and it didn’t happen.
So in the modern era of artificial intelligence – when machines can perceive and understand the world, when chatbots present themselves as eerily human, when trillion-dollar technology companies exploit powerful artificial intelligence systems to raise their advertising revenues – these recommendation engines are certainly becoming getting smarter, too. Normal?
Maybe not.
Recommendation engines are some of the earliest algorithms on the consumer web. They exploit a variety of filtering techniques to highlight the things you’re most likely to want to interact with – and in many cases, buy – online. If done well, they are helpful. In the early days of photo sharing like Flickr, a elementary algorithm ensured that the next time you logged in, you’d see the latest photos your friend had shared. Nowadays, advanced versions of these algorithms are being intensively implemented to keep you engaged and give their money back to the owners.
Over three years after reporting on what Pinterest had internally labeled a “miscarriage” problem, I’m depressed to say that my Pinterest suggestions are still terrible. In a strange way, Pinterest now labels me as a 60-70 year aged silver vixen looking for a stylish hairstyle. That and the sage green kitchen. Every day, like clockwork, I receive marketing emails from a social media company filled with photos suggesting that I might enjoy cosplaying as a coastal grandma.
AND was looking for #inspo paint online at some point. But I’m long past the painting phase, which just highlights that some recommendation engines can be intelligent, but not short-lived. They still don’t always know when the event has passed. Similarly, the suggestion that I would like to see “hairstyles for women over 60” is premature. (I’m a millennial.)
Pinterest has an explanation for these emails, which I’ll get to. However, it’s critical to note – so I’m not singling out just Pinterest, which over the last two years has put modern management in place and put more resources into refining the product so that people actually want to shop there – that this is happening on other platforms too.
Take Threads, a Meta company that collects much of the same user data as Facebook and Instagram. Threads is, by design, a completely different social media app than Pinterest. It’s a collection of mostly text updates, with an algorithmic “For You” tab and a “Following” tab. I actively open threads every day; I don’t come across it like I do when going from Google Image Search to Pinterest images. The Following tab in the Threads section shows me updates from journalists and techies I follow. In the “For You” tab, Threads thinks I’m menopausal.
Wait, what? I’m not a laboratory person. But over the last few months, Threads has convinced me that it does power To be. Right now, when I open the mobile app, I see posts about perimenopause; women in their 40s struggling to shrink their belly, regulate their nervous system, or treat late-onset ADHD; husbands who hire escorts; and Ali Wong’s latest stand-up special on divorce. This is Real Housewives-meets-elder-millennial-ennui a weird world that doesn’t quite reflect the accounts I choose to follow or my expressed interests.