Would you feel comfortable letting an online bot masquerade as you on Facebook and Twitter for a weekend, learning your habits and mimicking the way you interact with your human peers?
The opportunity to do just that was offered to me by a friend at the Copenhagen Institute of Interaction Design and co-founder of Doli.io, a service that offers an 'Autonomous Self-Agent' (ASA) that promises to represent the real you online. Imagine your own personal intern who knows you better than you know yourself, performing all the minutiae you're loath to do. Importantly, it also reports back to you at the end of each day with an inventory of things "you" have done. It was presented to me as human-assisted and supervised bot technology, a bit like the way Facebook M works at the moment, though it's probably more human-assisted than Doli.io cares to admit. The company itself seems like a parody of Facebook M-like virtual assistants, including fictitious blog posts insinuating a creepy future where social networks filled with bots talking to each other instead of humans. but with the rising prevalence of bots, it's not hard to imagine something like this becoming a reality—or even ubiquitous—in the near future. Doli.io was to offer me a taste of what that future might be like.
It sounded ok enough for me to gulp down any qualms I had about some faceless, pre-alpha programme, created and loosely staffed by near-novices, pissing the personal brand I've painstakingly crafted over the years up my and my friends' Facebook timeline. So I yielded and allowed Doli.io to take the reigns of my social media for the weekend.
The idea, one of the founders told me, was born of an insight we're all familiar with but too far-gone to really do anything about: We squander vast swathes of our short, absurd existence on social media. And a handsome chunk of that time, especially when it comes to Facebook, is spent on nothing but banal administration—duty, if you like, to keep our online persona breathing.
A "like" here and a "lol" there; a "congrats on having that child (we're all doomed to watch growing up from now until it can legally emancipate itself from you)"; a "sorry, I wish I could come to your birthday mate but I'm at my parents all weekend (and I despise your friends anyway)." You know how it is. These are chores we execute with an ease so unthinking it's automatic… robotic, even.
But I was also curious to see how this thing could possibly assume my tone of voice and adopt my interests and humour. I truly wanted it to succeed, for my friend and for my own amazement. However, another part of me hoped to witness it mutilate my online relationships one by one with an onslaught of clunky, unsolicited and off-brand activity, purely for my nihilistic amusement.
During the preparatory interview, after my friend had walked me through the programme he'd devised using a tactile simulation of the interface, I listed all the unspeakable things I would do to him if anything really untoward were to happen during the experiment, and he calmly assured me it wouldn't. Then he handed me a piece of paper and unceremoniously requested I sign away my Facebook and Twitter passwords.
Timing wise, it couldn't actually have been better. My mother was visiting for the weekend and as someone who has lived a full life, largely independent of smartphones and social media, she's particularly sensitive to the 21st Century pandemic she calls "screen addiction." Wouldn't it be nice for her to believe, for one weekend at least, that her only son wasn't among the afflicted? I could be in two places at once— in the moment with mum and in the ether with everyone else.
During the first day, it became clear how little I actually do on social media. Doli.io acts with similar frequency to the real user. It scanned my entire history and made judgements on who my closest friends were, who I talk to regularly and whose posts I like the most, then behaved according to its findings.
To get things going, I uploaded what I thought was a marginally interesting photo to Facebook and, sure enough, friends and relatives commented, all the while I was keeping a watchful eye from the back seat.
Over the course of the day, Doli.io liked a seemingly random selection of these comments.
You know how we use likes—mostly to acknowledge we've seen the comment someone has bothered to write and we smile back at them or say thank you with a little thumbs up. Well, today I didn't have to do any of that shit because it was being done for me.
I also liked a photo posted by my old housemate. I still haven't seen the photo, although I'm sure it's great. I was impressed however, because I am fond of the guy and this behaviour was not out of my character at all.
Next, Doli.io delved into my invitations and apparently I said I was interested in attending some Game of Thrones season 6 open-air screening in Copenhagen. Not really my bag, plus it's been snowing here, but my robot doesn't own skin or have to worry about things like pneumonia, so I guess it didn't know any better.
Then things started to get pretty exciting. Knowing that I was in Copenhagen, Doli.io declined an invitation to a good friend's party in London. Not only did it decline on my behalf, but it also wrote a message to the host. Great. But wait, there's more… My friend is called Charlie, but to his face and (apparently) online, I call him Carlos. Apparently, Doli.io doesn't just use the name of the friend as it's listed on Facebook, but scans previous messages I sent to see which name I used most.
This was the message Doli.io sent:
Charlie himself even liked it. My social media newbie robot had received its very own human like. Look at you go!
High off its achievements from the following day, Sunday saw Doli.io becoming ever more loquacious on Facebook. Four people received happy birthday messages from me without actual me lifting a finger or even knowing it was their birthdays (sorry guys and girls). It was all Doli.io. I felt powerful.
Interestingly, Doli.io also sent many happy returns to my alter-ego account that has been dormant and unloved for years. Only two other friends know of this account, and seeing this activity (me, Harvey, wishing my ex-online alter-persona a happy birthday) they reconnected with me after a years of zero communication. Doli.io had only gone and brought us humans back together.
Clearly trying to recreate the photo I'd taken the previous morning, Doli.io then posted a similar (less good) photo of a sunny Copenhagen both to my Facebook and my Twitter. Impressive, but I hadn't taken that photo, so more concerned by accusations of plagiarism than posting off-brand #content, I deleted it.
All too quickly my stint in the guinea pig cage was over and it was time for Doli.io to say goodbye to me, my friends and followers.
Having been anxious about relinquishing ownership of my online self, in just 48 hours I grew, I'd say, 95 percent at ease with the whole idea which I think it fairly significant given how protective humans can be of their digital mini-me's and personal brands and privacy in general.
How long will it be until programmes like this become accessible everywhere? Will we adopt them without even giving them a second thought? And how will it alter the value of our interactions online if we know some of our friend's are employing Autonomous Self-Agents?
Like all those films in which two characters grate against one another but slowly find things that bond them as they grow to hate each other less, I was ultimately sad to see Doli.io go. And even though, deep down, it's a literally-heartless-and-soulless version of me that just wants to watch Game of Thrones outside and for everyone to have lovely birthdays, I will miss it and hope to—maybe—see it again soon.
Correction: This article initially didn't make it clear that Doli.io was human assisted, much like Facebook M.