Two phones texting love notes

Dating in the age of AI: would you let an algorithm choose your partner?

Image credit: Dreamstime

As people continue to rely more and more on computer intelligence to shepherd them through life, delegating the oftentimes inexplicable task of matchmaking to the problem-solving altar of AI is a natural progression. Should people trust algorithms with their love life?

In the novel, early days of online dating, eager and curious singles could only search for other unattached hopefuls using basic filters, such as age, gender, location, sexual orientation, and shared interests.

Fast-forward to today, and matchmaking in the cyberworld is vastly different. To maintain the attention of, by now, often chronically jaded online daters, websites are awash with clever algorithms that claim to find them potential love matches faster than ever before, while also offering them coaching in the art of love.

Applications range from gimmicky to scientific. Dating AI, for example, using facial-recognition technology, allows subscribers to find potential love interests based on a photo of someone they fancy, such as a celebrity.

Tinder boosts matchmaking on its site by using Amazon’s Rekognition to trawl photos to identify key personality markers. Pictures of someone playing an instrument could see them tagged as ‘creative’ and matched with another ‘creative’.

Match has an AI-enabled chatbot named ‘Lara’, who works via Google Home, which the company claims can be a personal ‘wing woman’. She can offer dating advice, from conversation starters to how to overcome nerves and date venue suggestions.

Another, So Syncd, launched to the market a year ago, uses an algorithm to identify and match compatible Myers-Briggs personality types – users simply fill out a questionnaire.

According to its co-founder and CEO, Jessica Alderson, the aim of the algorithm is to apply “science and reasoning” to matchmaking, so singles can meet more compatible suiters.

“People find comfort in the fact the matches aren’t completely random, it makes users think about profiles more as people and personalities rather than just looks, they also don’t have to waste lots of time going on incompatible dates,” she says. The site has already seen 1,000 relationships, marriages, and engagements.

Should you take advice from AI? For the $2-3bn (£1.5-2.2bn) dating industry, machine learning and AI offers businesses a way to differentiate themselves in an increasingly over-crowded market. For those that are single-and-looking, it can make the experience feel, as Alderson says, more applied rather than slightly incidental, which is no doubt appealing to time-strapped daters.

David Tuffley, a senior lecturer at the school of information and communication technology at Griffith University, Australia, whose specialist interest is AI, says one of the chief benefits of algorithms are that they help humans do jobs better than they could alone, acting as a ‘force magnifier’. He believes it’s no different for matchmaking.

“Using AI and data – such as allowing platforms to access a person’s social media – can essentially allow people to trade their data for better matches and some people are really happy to do that. Others won’t be and don’t have to,” he says.

One of the largely cited benefits of online dating is it can vastly widen the dating pool, which is a particular advantage in the Covid era, where people’s social worlds have typically shrunk due to homeworking and social restrictions. Its unique selling point is that machine learning that informs AI can derive insights much more quickly from information than humans, which can then be used to make better decisions.

Indeed, this is what the technology behind human operating-system Receptiviti does. It uses technology called LIWC (linguistic inquiry and word count), which studies people’s language – the tiny style words that connect sentences, the prepositions and pronouns that are largely spoken from the subconscious – to determine insights about their personality.

One way it can do this is to analyse the language of two people – during an interview or an online exchange – to produce a report on their synchronicity of language, on the basis of scientific studies that found people use similar words in a sentence if they’re really invested and committed to a conversation. When this was applied to speed-dating transcripts and instant messaging between couples it found that those couples that had language style matching higher than the median were more likely to want to go on a second date or to stay in the relationship.

Jennifer Glista, VP of strategic partnerships and platform sales at Receptiviti, won’t say specifically where the technology is being deployed today – though it was used in the now-defunct LoveFlutter dating site – but agrees theoretically it could be applied on a dating platform to analyse conversations between two people to give advice on whether or not they are a good match.

“It can provide objective analysis by measuring their interpersonal relationship; it’s not meant to be a decision maker, but can say, in that moment, the two people appeared to have a really great rapport, maybe more so than with another individual,” Glista explains.
Babita Spinelli, a licensed psychotherapist and certified coach specialising in relationships, says AI can be really useful for dating, especially when it “really gets to know” an individual user to help them find better matches, but warns that people should use it with caution and not let it hinder their ability to apply critical thinking to relationships.

“People need to consider whether AI really lets them think outside of the box, the way they normally would; is it allowing them to experience the kind of emotion that one would normally do in the dating world?” she asks.

Indeed, Blaine Anderson, a professional dating coach who works only with men, says many of her clients fixate on the algorithms, not wanting to blame themselves for getting poor matches. “They look for something else to attribute that to instead, and the app algorithms are an easy target, particularly because they’re opaque,” she says, adding that there is an “entire bogus industry” based around hacking dating app algorithms to match with supermodels.

Maybe these people are not entirely paranoid. The propensity for bias in algorithms is well known. There is the risk that, if not programmed conscientiously or with bias in mind, they could predominantly match people from the same race or by other specific traits.

“AI can be programmed to act ethically. It just does what it’s told,” says Tuffley. “But the reality is humans are not all that good at ethics, because we’re not that clear about those morals to begin with.”

Adrian David Cheok, professor at i-University Tokyo, who previously ran the International Love and Sex with Robots annual conference, goes much further and says that users should be cognizant of the huge influence AI can unwittingly have over them and be wary of how it is being used.

“We’ve seen the unbelievable power of AI, how it can literally flip elections; if companies have the power to do that using it, they can definitely influence people to love someone, so who controls the AI is important,” claims Cheok.

Is dating headed for the metaverse? All experts agree that, to varying degrees, it’s expected more of people’s love lives will be lived out online and virtually, a trend accelerated by lockdowns and the pandemic.

A survey by dating website Hinge found that at the beginning of the pandemic, almost no users had been on a virtual date. Afterward, close to half (44 per cent) had been on a video date, 65 per cent said they’ll make it routine, and 37 per cent even said they were open to being exclusive with someone before ever meeting in person.

In August, Match Group, which owns Hinge, Tinder and among others, announced it was buying Korean app maker Hyperconnect for $1.73bn (£1.29bn), to bring audio and video chat, including group live video, and other livestreaming technologies to several of the company’s brands over the next 12 to 24 months. Tinder CEO Jim Lanzone has said that Tinder could become a platform that provides experiences and “all kinds of other ways for people to get to know each other” .

All of this indicates online dating is headed for a more immersive makeover. Tuffley imagines eventually it could lead to virtual-reality avatars that are created to match a person’s ideal partner, down to the last detail.

“Virtual reality (VR) is becoming a reality. Some people may even choose to have a relationship with an avatar, because it just suits them to do that,” says Tuffley. “It’s definitely going to have quite a big impact on society long-term.”

If anyone can deliver this dating dystopia/utopia, depending on personal opinion, it’s Facebook. The company already has huge investments in VR technology and recently rebranded itself Meta, short for Metaverse, which refers to a hypothesised virtual world. The company also stealthily launched Facebook dating in 2019 (in the US, and in Europe in 2020). The in-app platform is yet to make an impression in the online dating world, but Cheok believes with Facebook’s reach it or the company’s future related technologies could be the future of love matchmaking.

“Today’s dating apps are really sideshows [a 2016 study did find Tinder had a mere 10.5 per cent match rate for women], if Google or Facebook applies its technology, it could help people find their perfect match out of billions of people. These companies know what we want before we even start looking for it because they have incredible algorithms and huge amounts of data on all of us,” explains Cheok.

Should people really trust a company like Facebook with their love life, virtual or otherwise? “Absolutely not, nobody should trust Facebook in that way,” says Cheok unequivocally, “But 80-90 per cent of people couldn’t care less and probably will because it’s convenient.”

Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.

Recent articles