This week the group tutorial is about artificial intelligence, and the ideology baked into the code of the some of the tools and spaces we use.
mechanical turk was set up by Amazon as a space for humans to do small tasks that machines were struggling with.
The original “mechanical turk” was a chess machine from 1770, but of course the “machine” had a person inside.
http://blogs.scientificamerican.com/guilty-planet/files/2011/07/mechanicalturk.jpg
The modern day “mechanical turk” is human activity. Doing things that AI simply can’t do yet.
Chat bots.
Invisible Girlfriend might look like it is chat bots, but it was actually real people writing the messages. There is all the strangeness of an invisible girlfriend, the bizarre need to show off, or the desperately sad need for companionship etc.
So “invisible girlfriend” does generate many questions that could be explored but for this discussion it is the fact it was real humans who were ‘creatively writing’ the messages.
That the people being the “invisible girlfriend” (and “invisible boyfriend” too) didn’t send messages to same person they just logged in and then send to the next person on the list. The person answering had five minutes to:
- read the end-users’ profile
- read the last 10 messages
- then write a new response
If they are on minimum wage, that 5 minutes work would costs 55p. Apparently they were paid $.05 per message.
You can argue that platforms like Uber or Lyft are also very similar — a platform for micro jobs done by many people.
The invisible girlfriend example leads to huge areas for exploration — therapy, pretence, loneliness, status, etc etc but keep focused on the fact that a human is providing this serve and not a chat bot.
The human is doing the work that an artificial intelligence can’t do yet –— and it appears that often the human is just a tiny cog in a huge machine — are they just being exploited?
Another way of looking at this — it is obvious to say but important –— the people making the money are running the platform — the space for the micro work to happen.
cycle of capitalism: new machines do the work humans don’t want to do. Humans create new jobs in new areas of innovation for new types of work. Once this stage is fully achieved. New machines do the work that humans don’t want to do, etc…
It appears here that the human is no more than a clever machine part, their human value is of little interest — like putting a master chess player inside a box to fool the world.
The human isn’t doing something “creative”… just “repetitive” or “predictable”.
Andrew Keen’s new book The internet is not the answer tells the story of Larry and Sergei the guys who started google, he suggests that if you met them 12-13 years ago you would be confident that the free open internet create by Tim Berners-Lee was in safe hands. Larry an Sergi were libertarians in their outlook, passionate about openness and sharing and wanted to do no evil —– Keen argues that everything has now changed that the size and the control that comes with it are dangerous and needs regulating.
Fascinating questions to explore — along with the simple question — are we as all humans just mindlessly providing the “mechanical turk” type work for the huge platforms to ignore their responsibilities, make loads of money and crucially close down conversations?
Time to think through what are the implications of all this for art practice — other than to say… Is there a job to be done here? A job to –— reveal the ideology baked into the code?
comments powered by Disqus