As mentioned previously, Alan Turing is considered the father of AI. His theory of computation suggested that a machine, by using 1s and 0s, could simulate mathematical operations, and shortly thereafter, could simulate any formal reasoning. At the same time, discoveries in neurology, information theory, and cybernetics led to the first concept of building an electronic brain. Around 1956, the field of AI research emerged at the Dartmouth conference with John McCarthy from Stanford University. The first artificial intelligence program, Logic Theorist, was born. This program mimicked the problem-solving capabilities of humans by focusing on pattern recognition. This initiated tremendous research in the field of AI, leading people to believe that within 20 years’ time, machines would be capable of doing any work that people can do. At this time, various approaches were tested from programming knowledge into the computer to recognizing patterns through algorithms. Soon, progress slowed due to lack of tangible results, leading to an AI winter, causing funding of AI projects to dry up, further slowing the process to a near halt.
In the 1980s, AI research picked up again. Initial success of the so-called expert systems derived from emulating the decision making ability of humans, by extensively using if-then-else rules, acting on a growing pool of knowledge that was preloaded in the system, and then continuously expanded. Expert systems became the first commercial successes of artificial intelligence software.
However, public attention soon shifted towards the rising microcomputer, which put an end to early machines like the Lisp machine. The evolving new world of personal computing attracted most of the computing talent and caused a lack of attention on AI research. Soon a second AI winter arose, lasting until the early 1990s.
The personal computer was no solution in solving the challenges of artificial intelligence; with its standardized instruction sets, mostly serial in nature, the highly parallel nature of neural network computing advanced only slowly. However, the increasing graphics nature of personal computers, driven by ever realistic games full of special effects, demanded a highly parallel graphics chip architectures. New parallel architectures grew, ever decreasing in costs due to Moore’s law.
Due to this increasing computational power, AI progressed and gained greater emphasis on solving specific problems often in areas of logistics, data mining and medical diagnosis. This so-called narrow AI benefited from the growing computing power and the availability of more research experts in the field of artificial intelligence.
An area that greatly altered the perceptions of the general public was gaming. For example, the competitive area of computer chess-playing, which funded further development of IBM’s Deep Blue, becoming the first computer to beat the reigning world chess champion, Garry Kasparov on 11 May 1997. Thanks to advanced statistical techniques, large amounts of data being readily available, much faster computers as well as new machine learning techniques, AI continued to expand and flourish by the mid-2010s. Subsequently, IBM’s Watson beat the two greatest Jeopardy champions, Brad Rutter and Ken Jennings by a significant margin and Google AlphaGo’s won over world champion Lee Sedol in Go March 2016. People have since begun to accept that computers have made great advancements in artificial intelligence and a lot more is to be expected.
General public awareness has been stimulated and led to a general recognition of the inevitable evolution of AI. All of this was fuelled by events like the arrival of smart personal assistants on smartphones as well as success stories of AI progress, such as IBM’s Watson being used to help diagnose cancer. This mental shift towards AI also had an impact on the financial investment market, which has continued to heat up since 2015. The recent availability of affordable neural networks has also helped speed up the development.
For the past few years, significant progress has been made in concrete areas of computer vision, image recognition, natural language processing, pattern recognition, knowledge location, and machine learning subsequently leading to ever more applications of artificial intelligence in real-world problems. Over 550 startups alone, using AI as a core part of their products, raised $5B in funding in 2016. All leading technology companies have built capable teams of several thousands of people focussing on AI, with the additions of top research talent from leading institutions, which were often hired into the company’s teams. AI has become the hottest topic in the startup world, attracting the most funding for future oriented technologies in 2016.
The fast technological evolution of artificial intelligence in recent years will likely be the cause of a fundamental business transformation based on artificial intelligence. We are really only at the beginning. The transformation is fuelled by the impact of AI on software, the new capabilities in consumer products and the fast-paced advancement in robotics. The sheer scale of artificial intelligence investments, as well as the quickly growing number of people working in AI areas, lay the groundwork of a new economic revolution.