Sometimes humans get so efficient that it’s pointless. Cue, nuclear war. Could artificial intelligence (AI) be the next example?
Click HERE to see the visualisation in full screen, navigate within the window or pull the strings towards you if you want to explore all the nodes.
The AI game is not about making new minds so much as replacing old minds with machines. With this technology, every task - from marketing, to trading, to medical diagnosis, to writing scientific publications - is reduced to an algorithm that adapts based on experience. It is a machine that learns. Neural networks were invented in the 50's and the onslaught of big data has catalysed their utility.
Image recognition (see: DeepFace by Facebook to tell whether two photos shows the same guy in underpants) and speech recognition allow people to input data more naturally. Augmented reality allows people to receive the processed information consistently and non-invasively while carrying on normal life.
Depressed? Why You Shouldn't Talk To an AI Robot (Yet)
CleverBot - Siri's Brother, lets you chat with a robot about everything and anything
Cleverbot and Siri are talking machines, but the bots are still working on their conversation ettiquette. The catchy title of this post reflects our example below, in which we see that some AI in current form can dramatically misinterpret meaning. The subtleties of human empathy and humour are very difficult to program and are essential for AI to truely come 'alive'—as in the movie Her, which talked about falling in love with your robot.
AI is mainly being used for marketing and sales and the irony is that bots have started talking to bots. Will bots become superfluous to humans (or humans will become superfluous to bots)? What’s the point of robots marketing to each other? Imagine if Facebook becomes operated by machines - will it simply implode? The same with trading - the buying and selling could accelerate so much that we could have a system crash not every ten years, but every 20 minutes (See: 2010 Flash Crash).
What can't AI do?
And what about ‘consciousness’ - that ‘vital spark’ that differentiates humans from inanimate objects? What about the feeling of being ALIVE as a single entity? As Edsger Dijkstra, AI pioneer once said when asked whether a computer can think, it's a bit like asking 'whether submarines can swim'. At Google Calico they are ambitiously trying to erradicate death by programming a person. But are we just the sum of our knowledge and emotions? Where does the mysterious 'soul' fit into this puzzle?
What AI really can't do is experimental design - chosing the independent and dependent variables. IBM Watson cannot distinguish between 'input' and 'output'. Perhaps the next challenge will be to program politics?
Is this what we call breakthrough progress?
So AI can distinguish cats from non cats and predict that you will probably like Charlie Parker if you like Miles Davis. It's a start, but we've still got a long way before the stuff in the movies becomes reality. To get a program to carry out experimental design and political destiny we will have to pause and think about, well, the meaning of life (!). What are our overriding objectives? Not to mention AI's terrible sense of humour. Hopefully 2015 will provide a better picture of what AI can really offer as a benefit for the future.
nuviun's Datavision supports the stimulation of discussion and debate on important issues in global digital health. The views are solely those of the author(s).