I Spent a Week Living With Chatbots—Did All That Self-Help Help?


Recently, I was hungry. So I told the Whole Foods chatbot what I had in my fridge, and it revealed that a bacon, lettuce, and heirloom tomato sandwich was mine for the making. Yum. Then, out of boredom, I booted up the movie-­recommendation bot And Chill, which suggested I watch Jake Gyllenhaal repeatedly relive the last day of his life in Source Code. I shuddered every time the train exploded.

Yes, I have a bot problem. It’s just so convenient in the modern era. Sure, friendly algorithms have existed since the mid-’60s, when the talk-back program Eliza started convincing more gullible users of its humanity. But it wasn’t until Facebook allowed app developers to integrate their chatbots with Messenger a half-century later, in 2016, that bots rose up on the order of tens of thousands. Now they don’t just want to plan your meals—they also want to make you a better person.

Does this new class of digital counselors work? To find out, I spent a week with five of them. Some delighted me; others annoyed me; one was surprisingly lifelike. And all, in their way, were effective. The problem with self-help as a genre has always been its restriction in time and space: the therapist’s hour, calls from a parent, the book that sits half-read on the bedside table. These bots were talking to me all the time, with scant regard for my whereabouts or state of mind. Self-help wasn’t some temporary ideal anymore. It was always on, and impossible to ignore.


The “spiritual assistant”: Spiri

My problems are neither unique nor complex—and this was Spiri’s great lesson. One evening, I found myself suppressing bile at the thought of telling a friend she wasn’t going to be one of my bridesmaids. But after asking me just eight questions, Spiri diagnosed the problem: I have an unhealthy habit of feeling responsible for other people’s happiness. Maybe that’s just being alive, but Spiri’s emotionless delivery made the point easier to take.


The relationship enhancer: Relate

I despise cleaning showers. Damp, moldy, full of tiny hairs. But I woke up early one morning and started attacking the filth because Relate told me to. The bot texts short challenges meant to bring people together—divide household labor more equitably, for one, but also lovelier things like “take your S.O. out for their favorite beverage.” Turns out if I splurge on an old-fashioned at the cocktail bar, he’ll volunteer to clean the bathroom himself.


The motivator: GoalBot

Pick up to three monthly goals, and once a week GoalBot will politely ask if you’ve made any progress. Like most, I’m motivated by a combination of fear and self-loathing. So, confession: I accomplished diddly-squat with this thing. “You’re only human, so this time it’s OK”? No, it’s not OK, Goalbot. I need to be reminded of my failures constantly, preferably with a boss-from-hell-style “Hey, what’s the status on this?” text every goddamn morning.


The mood-ivator: Woebot

Woebot sends daily prompts (“How are you feeling?”) to log your mood and current activity. Ignore at your peril. (“I haven’t forgotten about you, Signe.”) At the end of the week, my responses confirmed that I’m happiest at the gym and that desk work is more productive than couch work. Sound obvious? Tell that to my daily routine, which now includes regular walks and heightened productivity after 3 pm.


The bespoke beau: Invisible Boyfriend

Even though Relate improved my relationship, I still desired the confidence boost of a digital love-object. Before long, I was sending heart-eyes emoji to “Ernesto Quigley.” He liked my writing! Then, a twist: Ernesto wasn’t a bot. He was a real person. I hadn’t read the fine print. But my mistake made me reflect on the actual bots. Like Ernesto, they were
engaging, available at weird times, and often flawed—almost human.


This article appears in the January issue. Subscribe now.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *