In March 2016, a computer prevailed over the world's best human go master for the first time. It was considered one more milestone on the long path to artificial intelligence. Even so, most experts agree that the path for AI to achieve human or super-human intelligence is still far off. Opinions diverge much more sharply about just how far away that moment is, and whether exponential development can be expected, as Tim Urban described in his influential posting on Wait but Why.
Philosopher Nick Bostom, a professor at Oxford, published a book in 2014 entitled Superintelligence, positing the much-debated thesis that AI will replace humans as the apex life form as soon as it achieves intellectual superiority. A great deal of fear is attached to the argument — fear of a potential, or even expected, superiority of technology. What is clear, however: the tech scene is wrestling with major questions about the nature of humanity.
In many ways, what's happening right now is just connecting the dots. Just as the internet achieved a new quality through the networking of many individual machines and Web 2.0 by networking many individual people, AI and technologies such as machine learning are engaging in new connections, such as between VR and AR to produce new interfaces between humans and machines. The distribution of roles between man and machine is being redefined in the process: the relationship between human-assisted AI and AI-augmented human resembles that of Augmented Reality (AR) and Virtual Reality (VR). As more human skills are assumed by machines and robots, the urgent question arises as to just what is the core of humanity. Is it solely those skills that cannot be automated? And what happens if none of those are left?
Artificial Intelligence (AI) has stood since its inception under an odd paradox: As soon as a function or skill once considered to be an element of AI is successfully implemented, it is immediately considered self-evident — and is no longer perceived as AI. One prominent example of this is speech recognition, once considered the holy grail of AI. Nowadays we talk to our phones without giving it a second thought, or even considering that it has anything to do with AI.
This paradox impacts the way men and machines interact. We interact intuitively with robots just like with other organisms, even when our behavior doesn't seem to make sense, given that robots only react to a highly limited degree — or not at all — to human behavior. Artificial Intelligence today is still highly reliant on human support. Max Levchin sees human-assisted AI as an inexorable trend. If Siri doesn't know it, then humans fill the gap. The fine line between man and machine is shifted, with the human now becoming the assistant to the machine. Those disturbed by that thought should remember that these scenarios have occurred in the past as well, such as the assembly line work of the industrial revolution. On the flip side, Levchin sees humans as returning more powerfully to the center of economy. Maximizing profit does not particularly benefit users — but they are the ones that ultimately decide economic success. Levchin names this trend 'beneficence': The doing of good – active kindness.
And this is where the circle is completed. Successful products put man at their center, not just the 'user' or 'consumer.' They serve mankind, not the reverse. Successful products have intelligent interfaces and offer people a new level of comfort and added value. A goal ultimately served by Artificial Intelligence and Machine Learning.