Stanford’s One Hundred Year Study on Artificial Intelligence report argues in second last paragraph on page 42 that artificially intelligent systems and humans have complementary abilities. It goes on further to claim that humans are likely to focus on tasks that machines cannot do as well, and gives the examples of complex reasoning and creative expression abilities of humans. While this may have been true in past, it may no longer hold true in next few years. Advanced algorithms are being developed capable of complex reasoning and even creativity, equivalent (if not better) to humans. A research paper on Creative Adversarial Networks (CAN) titled “CAN: Creative Adversarial Networks, Generating Art by Learning About Styles and Deviating from Style Norms” by Rutgers University in collaboration with Facebook Artificial Intelligence Research (FAIR) lab, and College of Charleston claims that machines are able to make painting as good as human artists and in some case people prefer painting created by artificially intelligent network to the ones made by human artists. The AI network was provided with 81,449 paintings from 1,119 artists covering wide variety of painting styles from fifteenth century to twentieth centuries. The research group conducted the experiment to examine if people are able to differentiate the paintings generated by machine from paintings created by artists and result showed that people could not distinguish machine generated paintings and paintings created by contemporary artists. This is not a one-off case and there have been multiple such events, for example another being “The Next Rembrandt” painting by system developed by Bas Korsten. Unfortunately, most of the current AI systems focus on mastering logic and not creativity. Creativity has been a neglected topic for a long time in AI research, mainly due to insufficient funding which in turn is because of muted potential of commercial opportunities in such endeavors.

Creativity is one aspect. There have been a lot of research going on understanding human emotions and researchers recently have been able to develop a system which can understand human emotions to a certain extent. It should not be too far in future when a robot can not only understand emotions but can also appropriately emulate them. Additionally, reinforcement learning and various other fields of AI are making AI systems smarter by every day. These AI systems are able to perform complex reasoning, one such example being AlphaGo developed by Alphabet (Google) which beat the game champion Lee Sedol in a five-game match of boardgame Go. It is evident from these events that AI systems and humans have a gap in their intelligence and they are not necessarily complimentary to each other. AI systems are not as smart or creative or emotionally intelligent as humans. Not Yet! But that does not mean they will never be. This gap between AI systems and humans can be attributed to the fact that research in AI has gained momentum only recently and research was not so active before 1960s. As research in this field deepens and the gap shrinks, human abilities might become a subset of AI systems’ abilities, and not necessarily complementary.
