“Robots and Weavrs” - David Bausola, Dan Catt and Meg Pickard at FutureEverything
Whilst I was at FutureEverything I did a video interview about the panel session I was on, and when asked what I hoped to get from the festival, I blurted out “Giant robots smashing up the city skyline” or something to that effect. The actual robot count at the event was very low, but a couple of talks did at least have robots in their titles or subject matter.
“We are not alone” - David Bausola
David Bausola of Philter Phactory was presenting their work with Weavrs, social media bots that had personalities without being true AI. The bots used the idea of being in a physical location to explore digital artifacts, and were given traits and interests to follow. A King Kong Weavr, for example, lives in New York and visits the Empire State Building every day, whilst the Clouseau bot lives, au naturellement, in Paris. They use social media to record their experiences.
They wouldn’t pass a classic Turing test, but then again, I’m not sure what percentage of my tweetstream would. David said that whilst they do not reveal that the Weavr’s are bots, they never describe them as human either. My colleague Dan Catt and I began to wonder whether you could set up Weavr’s for bits of media organisations or journalists who hadn’t particularly grasped the nettle of social media. It might be a useful carrot for the reluctant - unless you join in with social media, this robot is going to impersonate you badly!
Perhaps the best moment of David’s talk though was when he was asked “which was more important, content or design?”
“Use”, he replied.
“Robots, editors, strangers and friends” - Meg Pickard & Dan Catt
Robots were in the title of Dan Catt’s contribution to the day, as he was talking with Meg Pickard about “Robots, editors, strangers and friends”. Their talk was an explosion of ideas around the way that people collect and curate information and recommendations, and probably one of the only talks in which you’d see a breathless transition from vomiting pandas to downhill skiing to Guardian Zeitgeist.
The main thrust of their presentation was that people used to find out information and recommendations from a mix of “editors” who acted as publishing gatekeepers, and “friends”. To this mix the internet has added “strangers” and “robots”. The actions of strangers impact what is presented to us in the “most read” and “most shared” panels on websites, and robots and algorithms determine the type of content we see in functionality like “if you like this, you might like this too” or the aforementioned Guardian Zeitgeist.
Dan Catt also gave a great illustration of the law of unintended social software consequences. Whilst at Flickr the team had tried to detect news via the photographs that people uploaded. They worked out that, when something like a plane ditching in the Hudson happens, they would see a burst of photos all taken within seconds of each other in a single place. That may well work, but it turns out that exactly the same pattern of human behaviour happens when people get married. As Dan put it:
“Yahoo! invested a lot of resources in building the world’s biggest and fastest wedding detector.”