Why we’ve fallen out of love with dating apps

Dating apps have transformed how people meet romantic partners. But they seem to be falling out of favour.

Data shows that last year, four of the biggest dating apps in the UK lost over a million users between them. And research my colleague and I worked on suggested that this is because people have become frustrated and bored with digital matchmaking.

The frustration is usually the result of inconsiderate behaviour from other app users. And the boredom appears to stem from a growing belief that the AI used by these apps seems to prioritise short-term engagement metrics over meaningful connections.

This marks a significant shift from earlier online dating websites, which were notably more transparent about how they tried to establish authentic connections. Often this was through detailed answers to compatibility quizzes or personality assessments.

Get your news from actual experts, straight to your inbox. Sign up to our daily newsletter to receive all The Conversation UK’s latest coverage of news and research, from politics and business to the arts and sciences.

For example, OkCupid.com (founded 2004) asked users a wide range of multiple-choice questions. It then went further by also asking them to specify the responses to those same questions they wanted to see from prospective partners.

In contrast, today’s dating apps increasingly rely on less transparent AI which seems to be based on simplistic engagement metrics (number of swipes, frequency of texts, time spent on the app) rather than a path to genuine compatibility.

The result is often a selection of vague, fleeting connections that do not amount to meaningful relationships. And the business model of today’s dating apps – selling increased visibility and access to matches – creates a challenging environment for many users to find the matches they want.

Because of this, many users experience a sense of dissatisfaction which manifests itself in four stages.

It begins with what I call a “boredom cycle”. General boredom prompts many daters to use the app in the first place, but resulting conversations often turn into uninspired and lacklustre exchanges. This adds to the boredom, which then spreads and grows throughout the entire dating app network.

After that is a general sense of disappointment, as users become jaded from regular “ghosting” (cutting off communication without notice), “flaking” (cancelling dates at the last minute), and mundane message exchanges which don’t lead to actual dates.

This all leads to a third stage of “algorithmic cynicism”. At this point, users become increasingly sceptical of dating app algorithms, suspecting that their primary function is to encourage the purchase of certain features rather than to establish authentic connections.

Finally, communication fatigue kicks in. Users go through the motions of swiping and texting with a sense that there are no better alternatives. It all becomes a somewhat hollow experience which ultimately drives many away from the platforms completely.

Swipe on, swipe off

Research has also shown that the initial rise of online dating usage among millennials coincided with early enthusiasm about social media. But this enthusiasm has diminished.

Social media users are now increasingly suspicious (and vigilant) about the risks of misinformation, scams, and offensive content.

Despite all of this, people still seek connections through dating platforms – whether for casual or long-term partnerships. So perhaps the issue is not with digital dating itself, but with how the industry uses AI. And an alternative is possible.

In related research on human-AI joint decision-making, my colleagues and I found that when people interact with an imperfect AI system, but also receive clear feedback about the their own behaviour patterns and how the AI responds, they can help to correct errors.

That research focused on financial decisions, but dating apps could do a similar thing by openly providing daters with personalised insights about how AI algorithms are responding to their activity. This is something that no dating app currently does.

Instead, they let users adjust certain filters, such as age, location and ethnicity. But then they use AI to create “revealed preferences” based on people’s patterns of engagement with the app, like swiping and messaging. These “revealed preferences” seem to greatly influence the kinds of profiles that the dating apps’ AI recommends.

For example, even if someone says they are keen to date people across a wide age range, the app may still end up recommending profiles of a narrower age group, because the user has tended to swipe right on those in the past. Because of this, users have expressed concerns that the AI could be adding unwanted limitations to their potential dating pool.

In the AI of the beholder

Indeed, our research shows that a cynical view of dating app algorithms is a key reason why something as potentially exciting as finding a romantic partner can become so dreadfully boring.

Addressing this issue, by simply explaining to daters how AI interprets their use of the app (the swipes, the matches, and actual dates) could be a valuable selling point. Giving users freedom to adjust other filters besides demographics, such as those related to their values and interests, could further increase interest.

This would represent a return to the more transparent match-making principles of earlier dating websites, but with the benefits of the latest technology.

Our research suggests that as dating app users grow ever more discerning, they will demand greater transparency and an improved overall dating app experience. The industry’s future may ultimately depend on whether companies can shift focus from impersonal engagement metrics to fostering authentic connections.

And platforms which embrace transparency and empower users could make many fall in love with dating apps all over again.

Anh Luong does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.