You may remember a few weeks ago I wrote a startling and delightfully informative and witty piece about LivesOn, the experimental service which promises that when your heart stops beating, you won't stop tweeting. You can read that thing HERE, but if you're too lazy then let me summarise it;
LivesOn.org analyzes your Twitter feed and, over time, learns to mimic your posting style and personality so that when you inevitably die it can simply slide into your place and keep you alive via social media.
Horrifying and brilliant all at once isn't it?
LivesOn.org is powered by geniuses at Queen Mary's University and thought up by Lean Mean Fighting Machine; a London based ad agency. I got hold of Dave Bedwood, the acceptable face of virtual poltergeists, to get some questions answered...
First and foremost, what do you do sir?
I am a founder and creative director of an ad agency called Lean Mean Fighting Machine.
What made you think to invent the digital ghost? Was it that one episode of Black Mirror?
The idea was born in March 2011. As an ad agency and spend our lives thinking about creative ways to use technology and social media. You realise how much information about themselves people are putting on to these platforms.
We got to thinking about the social media of the future and, with tech becoming ever more pervasive, how much of ourselves will we put online – effectively our ‘zeros and ones exist on servers all over the world. We’re interested in A.I, and automation so we thought about whether technology could learn about us well enough that it could in the future create online version of ourselves.
Most people these days are just sticking pictures of what they eat online. Which doesn’t really reflect a complete person. Will any of what we post really constitute ‘us’?
But it's hard to see this trend reversing. With tech like google glass, we will give more and more of ourselves to social media. And if we are posting all this stuff what will happen to it?
It feels evolutionary in a way, inevitable that people will use technology to somehow live on. Popular culture has explored this for years - from Lawnmower man to the Matrix. Religions have sold us afterlife, good or bad, for eons.
Anyway, two years after that initial thought, a planner here read an interview with Charlie Brooker, the writer of Black Mirror. He didn’t give away any story lines, but did say an upcoming episode would feature social media and death. So we thought what perfect timing. So we created a site in about 1hr, worked out roughly how it could work on twitter, contacted Queen Mary University (we do a lot of R&D projects with them) who were interested in working on the project.
However, we hadn’t actually made anything, it was just a concept. We tweeted the link to our site when the black mirror episode aired. Then it just spiraled out of control.
Was tweeting after death the original intention for LivesOn?
Well, that was the original thought. But actually as we talked about it we felt it could actually be just a very good aggregation tool. Like Netfilx or Amazon. It would learn about you, your likes, dislikes and sort of become an online twin. So you wouldn’t have to be a slave to your computer or phone, you won’t miss anything as your ‘twin’ would see it and save it for you.
But the death angle was just a far more potent a story.
How does LivesOn work?
Currently it looks at your syntax and sentence structure in your tweets and then uses Markov chains. Which basically means it looks at the chain of previous words and from that predicts what the most likely next word would be. With some randomness thrown in.
So it gives pretty funny, sometime accurate results. We will at some point work on the aggregation tool, and hopefully make liveson not just a weird, funny, thought provoking thing but also a useful tool.
I've signed up. I think I was member 189. Some of the tweets LivesOn produces are very nearly spot on, whereas others are a sort of mad patchwork. How does LivesOn train to better emulate a user?
Well, we’ve limited the number of people at any one time, as this is only an R&D project we don’t want it to fall over or cost us too much money to run.
In terms of emulation, that’s part of its charm, it is sometimes spot on, mostly odd, a bit random. You can favourite tweets that you liked or found funny, or close to how you speak and it will get better. But, it is nowhere near true A.I. That is cutting edge computer science, this is more about a provocative idea that hopefully makes people think, as proper technologies and A.I will most likely happen at some point.
LivesOn requires you to appoint an "executor" of your virtual will. What happens if you both die?
Good question. I don’t know. We would presume that most people have some form of network of friends, especially ones that are on social networks already. And if this is the case and we are tweeting from this persons account (remember all accounts are private anyway unless the user changes that) someone from their community would find it easy to find us and we would just cancel the account.
What's your ultimate goal for LivesOn? Are there plans to link it to other social networks, or have it ring your family late in the night?
No, it was an R&D project, not a business start up, so we will either invest in it a little for the aggregation development, or just leave it running as it is now. But there are no grand plans. It has given us quite a lot already.
Words by Gazz Wood
