Article Thumbnail

Breaking Up Is Hard, But What If You Had an A.I. Do It for You?

Researchers are starting to train artificial-intelligence systems to sense relationship red flags using invasive phone surveillance. But if you’re okay with being watched by a machine, it might be able to tell you whether someone is a bad match long before you ever realize it

Bizarrely, I’m no stranger to using artificial intelligence for relationship purposes. Back in 2019, as part of a very important journalistic investigation, I used two A.I. apps to find out how well I romantically matched with my friends, exes or my now-boyfriend, and to decipher which contacts in my phone fancied me. Spoiler: Literally everyone fancied me (except my mum, thank god). So, since A.I. correctly predicted my compatibility with my boyfriend, could it one day help me figure out if I need to break up with him?

Luckily for him (or unluckily, maybe), not only do I quite like him, but the technology doesn’t actually exist yet. Still, that doesn’t mean people aren’t already thinking about how effective it would be if it did. This is exactly the question that customer service platform Tidio posed to the respondents of its most recent study, “Love in the Age of A.I. Dating Apps.” Unsurprisingly, when asked what they would do if an artificial intelligence dating app did suggest that they break up with someone, only seven percent of people said they would end the relationship immediately. Forty-eight percent of the respondents said they wouldn’t, but would keep an eye out for red flags following this advice, while 45 percent said they wouldn’t take any advice from A.I. if they really liked the person they were dating. Notably, these figures only apply to short-term relationships formed on a hypothetical A.I. dating app, not long-term relationships or relationships formed offline. 

How would A.I. even be able to offer this kind of advice, though? What kind of information would it need to be fed to come to the conclusion that your relationship was doomed? “A.I. dating apps of the future could use information about a person to create a set of parameters, which the recommendations would be based on,” says the study’s author Gosia Szaniawska-Schiavo. “For example, the engines could use a person’s historical data to predict the likelihood of the relationship being potentially successful. This historical data could be in-app user behavior, such as historical matches, conversations that had gone hiatus and changes in the conversations (based on keywords and frequency of the conversation) that the A.I. could identify as a change of heart and, consequently, a potential threat to the relationship.” 

For this to happen, the A.I. would have to be constantly mining your messaging history and romantic preferences for data. 

When it comes to long-term relationships or those formed outside of an app setting, it’s less clear how A.I. could ever offer breakup advice. The main issue seems to be how it could possibly be trained to make this kind of suggestion. It could use data from text exchanges or rely on the person asking it for advice giving it information about the relationship — but neither of these are particularly reliable methods of getting the full picture of a relationship, and therefore being able to advise whether it should end.

“The A.I. would need to be very advanced,” says 27-year-old Yasmina from England. “You couldn’t relay your problems to Siri and get an answer. It would also need to be balanced, because you’d only tell it the bad stuff. Maybe one day when A.I. listens to all your calls and reads all your texts — and therefore knows your life — it could provide something balanced.” Although she wouldn’t take the advice, Yasmina does say she would ask the A.I. “out of curiosity” if she was thinking about breaking up with her partner, but would also “talk to real people.” “It could maybe plant a seed in your head, though,” she adds.

One redditor did come up with a possible training system for the A.I. when I posed this question in the r/ArtificialIntelligence subreddit, though. “There’s a psychologist, John Gottman, who studies relationships, and he can supposedly predict the long-term success of a relationship based on how the couple fights,” wrote Ella_surf (seemingly an A.I. fan, as opposed to an expert). “Certain behaviors (he calls them the ‘Four Horsemen’) show that a couple won’t last. So you could potentially use recordings of lots of couples fighting, see which of them are still together 10 years later and then train an algorithm on this. If you combined your vocal text analysis with some intonation analysis, and maybe even video/motion capture for some body language analysis, I wouldn’t be surprised if you could get some decent predictive power.”

While this is a pretty decent suggestion, training an A.I. on something like this would not only take years, but is a logistical nightmare. Either these couples would have to be recording themselves all the time, or they’d have to stop mid-argument to set a recording device going. Also — obviously — each individual person and relationship is different, so it seems unlikely (see: impossible) that one system could work on every romantic relationship in a specific location or culture. There’s also the incorrect assumption that all couples fight, or fight the same — some couples hardly fight, but that doesn’t mean their relationship is healthy, or that they should stay together. Thus, fighting wouldn’t be a reliable indicator for everyone. 

However, when based solely on interactions via an A.I. dating app, and used only by those who met on that app, artificial intelligent breakup advice might be more reliable — though still not worth trusting completely. “This whole approach to classifying and predicting people’s behavior is based on the assumption that love — which is a strong emotion in itself — could be found based on rules and logic,” says Szaniawska-Schiavo. “To put it simply — the A.I. will be 100 percent trustworthy only in its own paradigm, which is purely based on ‘if this, then that’ logic. For example, if the A.I. would be programmed to recommend a breakup because our [dating app] match, let’s call them person X, already connected to similar people like us, and then stopped having conversations with them after a few meetings — would that be a trustworthy recommendation for us? It could be seen as logical. However, we have to remember that finding love is not always logical.”

This is the view held by most of the redditors who responded to my question. “Not trustworthy at all,” wrote It_Matthew of A.I.’s breakup advice. “Sure it can detect signs of common big flags. But A.I. can really only work with things that are consistent, and humans aren’t consistent.”

London-based Alexandra, 24, is a little more forgiving of the idea. “It depends how much A.I. could understand about my relationship,” she tells me. “I’m a massive sheep, so I really take advice from anyone. If A.I. could grasp the deep ins and outs of my relationship, then I’d definitely take advice from it — so I’d stop having to ask my friends.” Alexandra adds that if someone is using an A.I. app to decide if they want to break up, they’ve probably already made the decision themselves. “If you’re in the stage of wanting to break up with someone, you try and find any excuse [to do it], so maybe A.I. would be good for that,” she concludes.

Szaniawska-Schiavo sees using A.I. for this purpose as a simultaneously good and bad idea. “It might be a good idea because the A.I. would remove the emotional factor, and focus on the actual potential of the relationship based on specific factors such as behavior, interests, goals, etc.,” she tells me. In a dating app setting, she adds, “this could put us closer to people we want to meet straight away and remove the mystery from the process, which could be useful to those who don’t want to repeat it one too many times. With A.I., the screening would have already been done for us.”

But as most of the redditors who commented on my post also said, and Szaniawska-Schiavo repeats, “finding love is not always logical.” “The A.I. would be calculative, and its recommendations would be based on rules,” she explains. “But emotions aren’t based on calculations and removing risks — very often, finding love stems from taking them. Let’s say you meet someone who is good for you ‘on paper,’ but you don’t really like them — was meeting them a bad idea in the first place? Well, it depends on how to look at it. Maybe you just realized that you like people who are your opposites when it comes to temperament. Or maybe, you noticed one characteristic that this person didn’t have, and you just realized that without it, you can’t commit. People are very complex when it comes to emotions, and this is something A.I. can’t fully grasp — for now.”

If artificial intelligence does one day become sophisticated enough to offer valuable and reliable relationship advice, at least I could break up with my boyfriend using the exonerating excuse: “It’s not you, it’s the artificial intelligence.”