Many of us interact daily with artificial intelligence starting with personal assistants Siri and "OK Google" on smartphones, social networks that offer you to friends of those who sometimes you forgot that you knew and the usual vacuum cleaners, which assess the level of contamination and obstacles in your house...
Deep learning AI systems, allows you to connect not only people on the "scraps" of information, but also to calculate, to predict events, possible risks and deviations from the standing targets in real-time. Thus, when the stock trading for many years used learning algorithms of artificial intelligence.
As you know, in some countries we can safely add to the list of AI self-driving cars, logistics companies on the task parameters, a variety of drones, amazing results build medical predictions for the settings part (which experts and doctors are not able to assess due to the huge volume of information about patients from around the world) and much more...
According to analysts, only with the introduction of AI into systems of relations with clients (CRM) and fraud risk assessment within 3 years artificial intelligence will allow to increase revenues by more than a trillion dollars.
We must not forget about the various smart devices, the so-called "Internet of things", which creates a base for subsequent learning of artificial intelligence, because by and large almost all subjects potentially have an interest in passing any data for analysis and forecasting (starting with smart house systems, and ending with off-grid power plants).
So, after the inception of big data Analytics with AI in large companies has become much easier to work and make decisions with fewer errors and human factor. And so released resources will be used for operational decision and other tasks.
Modern supercomputers can of course solve the problem, but they rely solely on "brute power", and evaluate the deficiency for example, IBM has announced the creation of the US army computer operating according to the principles of the human brain.
Its performance is equivalent to the system of 64 million neurons with billions of synapses 16 (related to the type of the brain). In this case, each core will be part of a distributed network and work in parallel with others (even if one of the cores).
It's amazing how erroneous are the estimates, even 15 years ago reading the views of futurologists, I, like many suggested the development of artificial intelligence and automation of simple, trivial things. It seemed to me that the AI be replaced by a simple profession, which do not require high level of knowledge, and now I see bots with the qualitative algorithm is easy to substitute lawyers, and data Analytics replaces a medical specialty. To replace simple functions in the long term is not profitable to invest, to pay wages to the person much cheaper. Yes, I spent enough time studying the works of Alan Turing and others.
It remains to hope and believe that the tendency of many countries to address the likely surge of massive unemployment, will come to the realization already described approaches provide a universal basic income (when all will be paid the minimum required amount of money). Examples from history tells me that this is pure utopia (to spread around).
It turns out that were only prepared to deal with the ethical and moral issues in the development of artificial intelligence for the benefit of all.