The basic concepts of AI have not changed much since Alan Turing’s time, however the approach to achieving it has. Significant progress has been made in the last 5–10 years due to advances in machine learning and deep neural networks, leading to a range of services (some shown above) that we all can readily experience today. Like any other technology, AI will evolve from addressing simple tasks to vastly more complex problems. The evolution of AI is commonly modeled in 3 stages. We are currently at the very beginning of the first stage.
1- Artificial Narrow Intelligence (ANI): In this stage, AI is focused on one specific area like translation, face recognition, playing chess or diagnosing cancer. Here the AI has been trained to support decision-making in one, and only one, area. An image recognition AI would have no clue about a game of chess. ANI could be compared to hiring a top specialist for doing just one job better than anyone else—?but is fairly useless for any other tasks. Currently (2017), all AI applications are at the ANI stage. In the next 3–5 years, many new products and services will integrate ANI features offering unique customer value?, creating ?a huge business opportunity for differentiation.
2- Artificial General Intelligence (AGI): The next stage is AGI, where the AI system is as intelligent as humans across the board performing any intellectual task that a human being can at a comparable level of expertise. An adult human acquires an abundance of knowledge and variety of skills and is capable of using all these in many combinations. Matching this is a lot bigger challenge than achieving expertise in one area. As of yet, no one has achieved creating an AGI. Average prediction by AI experts for achieving AGI is around the middle of this century. More optimistic predictions are by the year 2030. “When Will AI Exceed Human Performance? Evidence from AI Experts (https://arxiv.org/pdf/1705.08807v1.pdf)”
3- Artificial Super-Intelligence (ASI): AI expert Nick Bostrom defines Super-intelligence as “an intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom, and social skills.” Much better can mean anything – 10x, 1000x, or 1000000x. It is humanly impossible to imagine it, and even more, its impact on our lives and society. Experts prediction for achieving ASI is sometime in the second half of this century. When intelligent machines themselves start designing their next generation versions, progress becomes exponential and unimaginably fast. ASI is the stuff of science fiction at the moment.
Currently, and for the next few decades, all AI will be in the narrow segment (ANI). With this said, we should not underestimate what can be achieved by combining various automated AI decisions for a business solution. As an example, if video streams from all surveillance cameras in a public place are fed into a machine learning AI system, over time the system is able to establish the pattern of video data for a normal situation by itself. The system can then detect an exceptional situation based on the shifting patterns in the video feeds. Thus, based on the situation it is able to recommend the best action based on previous situations and decisions.
In this book, we will focus entirely on ANI technology and applications because it offers vast and immediate opportunities for business innovation in almost all industries. AGI and ASI are way beyond the scope of most businesses in this decade.