Would you let an algorithm manage your relationships?

By Tim Maughan

There are few aspects of our lives that are not influenced by algorithms. But would you let one manage your relationships with other humans?

“I had this everyday feeling – stress about not properly articulating my emotions in my emails to people,” artist and writer Joanne McNeil tells me over the phone from Boston. “I was feeling as though I had to over-do it with enthusiasm or I would sound too sarcastic or bleak or disinterested.”

It’s a common anxiety of modern day life: in an age where we increasingly communicate via email, text messages, and social media posts instead of face-to-face, it can be hard to judge whether we are getting the tone right. Are we being too formal? Are we being too familiar? Are we unintentionally coming across as angry or unfriendly? Without the non-verbal cues we take for granted when talking in person with someone – or even on the phone – it can be hard to know whether what you’re saying is being taken the right way.

But what if there was an app for that? What if you could hand some of that responsibility over to an algorithm that calculated what to write in an email? Would you trust a piece of software to communicate with your boss or your loved one for you? Or – going even further – would you let it advise you on what to say when you were on a date, or tell you which of your friends you should hang out with, and which you should avoid?

It was while thinking over these issues that McNeil came up for the idea of Emotional Labor, a plugin for Gmail that scans your messages and inserts overly-familiar, lighthearted touches to make them seem more friendly. Full stops become multiple exclamation marks, ‘lols’ are dropped into sentences, and formal sign-offs become rows of kisses.

“It's over-the-top. It's completely over-the-top,” explains McNeil. “I could have easily had it just swap out periods for exclamation marks, but it swaps out one period for multiple exclamation marks. I could have used less saccharine and treacly words but I used words like hearts and stars, and made it really overwhelmingly crazy friendly. The idea came first from imagining this dystopian future where email would be written for you by some automated software that would perfectly articulate a friendly tone of voice. I was building something that works and does the job but also makes fun of the possibility of that kind of app existing.”

McNeil’s app seemed to hit a nerve when she released it online, getting shared virally across social media and, of course, by email. “I didn't expect that would happen. I didn't realise how many people also have that stress in email communication. Not just emailing friends, but emailing a coworker and being afraid that you're not sympathetic enough to someone who is, say, sick from work and you want to express that you wish they're healthier soon. There aren’t many non-cliched ways of expressing that, or expressing concern or caring for someone. We only have really limited phrases and a lot of them are stock phrases.”

While Emotional Labor was clearly a tongue-in-cheek art piece, McNeil believes people might be ready for some assistance with these kinds of tricky communication tasks. “That experience of anxiety, of not knowing what words are right – it’s not just that we might outsource it to algorithms. We sometimes outsource it to your friends because this is just a constant problem. How do we come up with the right words? For example, when you meet someone and you like that person, people often have friends ‘workshop’ their texts. If you're trying to set up a date with someone you might reach out to a bunch of your friends and say, ‘What should I say to this person? How should I sound flirty? How should I not be too overt with a come on but also sound interested?’”

This idea of crowd-sourcing advice on how to handle a date or a personal interaction is also a common theme in the work of Lauren McCarthy, an artist, software developer, and assistant professor at UCLA’s Design Media Arts programme. Her smartphone app Crowdpilot allows you to surreptitiously stream a conversation – say, on a tricky first date – from your phone to a group of other people online. Some of them might be your friends, whilst some are just randomly selected, anonymous Crowdpilot users. While listening in they can use the app to give you tips on what to say, suggest topics of conversation, or tell you how to react to the other person.

There’s a wonderfully cheerful, realistic advert-like video that explains how it might work. Although, like Emotional Labor, intended as a work of art, the app was also available to download on the App Store. “It was important to us to build a functioning app, so that it would go beyond speculative design fiction or sci-fi,” McCarthy says. “Because it is a real app, when you encounter it, you are faced with choices and questions. Will you download it? Will you use it? Maybe you find it terrifying or dystopic, but what happens if it actually improves your life?”

“While my work deals with technology, it is at its core dealing with being a person in modern society, which happens to involve a lot of technology. I am most interested in the performative aspects of social life and how we navigate relationships and interactions. Adding technologies into that, making apps or pseudo-startups or devices, offers another way to explore this.”

“I’ve always felt like I was sort of awkward and socially inept, and I initially wondered if I could build technologies that would help me out with this,” she adds. “This investigation started with a hat that would detect if I was smiling, and stab me in the back of the head if I stopped, in order to condition my brain to smile all the time. What I realized through doing this work was that there was always this element of failure and dystopia – I became really critical of the potential for tech to solve all our problems and realised it actually created a lot of new ones, too.”

Moving on from crowdsourcing relationship advice from other humans, McCarthy started looking at how the technology itself might start offering guidance. Along with her partner Kyle McDonald, she developed Us+, a rather startling app that monitors your video chat conversations and gives direct instructions on what to say and do.

“I had been doing a lot of research into linguistic analysis and Kyle had been doing a lot of computer vision research, and we wondered what would happen if it was applied in real time to a conversation,” she says.

The app analyses both what you say and your facial expressions while you talk, and will give you feedback on your performance. It’ll tell you if you’re being too self-absorbed or aggressive, suggest you be more positive or sympathetic, and tell you if the other person looks sad or happy. It’ll even mute you altogether if it thinks you’re talking too much.

Again, there’s another clever video that explains how it works. It feels very much like creepy science fiction, or something from an episode of the TV show Black Mirror, but McCarthy says they had surprisingly positive reactions to the project when it was released. “We meant to pose a question – do we want a future where humans have their every word and expression and reaction monitored and augmented by technology? We were critical of this idea, but at the same time, there were a lot of people that got in touch interested in Us+ as a real product – a business solution, self-help tool, or relationship improvement app. Even when we explained it was an art project, they didn’t really care, they still wanted it.”

Another McCarthy and McDonald art project that’s also a real world app is pplkpr. Designed to ‘optimise your social life’ (according to itspromo video), the app combines GPS data from your smartphone with heart rate data from a smart watch, or Fitbit-type wearable device, to work out when you are meeting people and how you are reacting emotionally to them. An algorithm then crunches this data to report back to you on which people you should be hanging out with more, and which you should avoid. It knows who makes you happy and excited and will send them texts asking them to hang out with you more, while it’ll even delete the contact details of people that make you bored or angry.

“When the app was first released online, there was a huge range of reactions and a lot of loud debate,” says McCarthy. “Some people were outraged – they felt this represented technology going too far and the end of humanity. Others thought it seemed practical and useful and wanted to try it themselves.”

“We heard from VCs that wanted to fund us, doctors and therapists that wanted to try it with their patients, researchers that wanted to collaborate and share findings, start-ups that were working on similar ideas and wanted to hire us, and people that felt they just really wanted and needed it. The point was to provoke discussion and thought, and I think that can happen regardless of whether you know it's an art project or not.”

You can still grab pplkpr from the App Store, so I installed it and tried it for a week as I ran around to meetings here in New York. I don’t have a Fitbit or a smart watch, so instead the app constantly nagged me via notifications about how I felt. It seemed to me like the app was transferring one kind of work – emotional labour – into another; it became digital labour, the endless tasks that our so-called time-saving devices end up making us do.

In all I found the experience rather depressing, but I wondered if that was perhaps me projecting my own fears, especially as I might be older than many of its users. “Younger people often find it much less depressing and are willing to engage and try it,” McCarthy says. “When we tried this with undergrads we realized how much more open to different technologies and interactions they were. They didn't come in with preconceptions or fear.

“They are not afraid. They are willing to understand something before they judge it, and hopefully this will mean a future where we can openly debate changes and new tools, and play an active role in building the world we want to live in.”

What’s interesting is that while both McNeil’s and McCarthy’s art pieces are forms of speculative design, since they were first released an increasing number of ‘real’ similar apps have been developed, apparently aimed at more serious users.

One particular example is Crystal Knows, an app that taps into your LinkedIn account and advises you in how to write emails when dealing with business clients, in a similar – if less fun – way to Emotional Labor. But McNeil is doubtful about how seriously they’re used. “I haven't met a single person that uses things like Crystal Knows on a regular basis. It’s too much of an experience that feels inauthentic or like you're cheating. [These apps] are out there and they've been talked about but it doesn't appear that anybody is really incorporating them in their work. I think it's just because even if you're unsure about the words you put in an email they're still your words. They weren't offered by someone else.”

Not that she rules these apps out completely, and offers an interesting comparison as to why they might appeal to some people even when they know they’re offering an inauthentic experience. “I do think it’s like astrology in this sense where, not everybody, but plenty of people know that it's fake and yet still read it because there's something comforting about having answers.”

And maybe that’s the real answer as to why some of us might use apps and algorithms to advise us on tricky personal relationships: not because we truly think it works, but that it gives us a little bit of hope. Amongst all the rationalisation of our technologically regimented lives, we still want to believe that something might support us, allow us to off-load our self-doubt, and give us the answers we don’t have. Whether it’s advice from our friends, horoscopes, or a smartphone app – maybe in uncertain times we all just want to believe in something more than ourselves.

/By BBC/