AI in Our Eyes

Are you worried by artificial intelligence (AI) or are you hopeful for what the future holds? AI has long penetrated people's lives. For example, the subway in Guangzhou, Guangdong Province, China has essentially achieved driver-less driving. There's no doubt that AI is going to change the face of employment as we know it. What's more frightening is whether we will become slaves of artificial intelligence? In 1968, in the opening of the sci-fi movie "2001: A Space Odyssey" directed by one of the most influential directors in cinematic history, Stanley Kubrick, it suggested such a serious issue of whether AI would betray humans. In the movie, the AI spacecraft, Hal, rebelled and even killed the human operator. So what exactly is artificial intelligence and how will it impact on people's lives?

Black Box and IterationWe believe the fundamentals of artificial intelligence is the black box and iteration. A so-called black box is something the outside world is unable to open and the operation process is unknown. The computer is like a big black box, and most of us do not know how it works even those who specialise in computer sciences. Nevertheless, we find the modern society that even a seven years old child can use smart devices without an issue. This is the black box theory. The black box features an extremely complex internal structure but an extremely simple external interface. Users do not need to understand the operation process behind, as it's enough to know that a specific input results in a corresponding output. Those smartest and the most specialised expert often package a solution into a black box. No matter how complex this solution is, the outside world can easily access and transfer information.

Iteration refers to the repetition of procedure applied to the result of previous applications as a means of obtaining successively closer approximations to the most precise solution of the problem. According to the famously known Moore's Law, the calculation speed of computers have continued to double every so often, feeding into the rapid grow of computing speed in smart devices. The power of high computing speed is the foundation of iteration, making possible faster and faster iteration.

What is intelligent about AI?

The most intelligent area of AI lies in it's self-learning ability, so called machine learning. Machine learning is essentially getting computers to replace humans in running codes and using iteration to continuously improve operations until an optimal is reached. Machine learning is 'slow', thus must be 'trained' using vast amounts of real life data and statistics. Big data and cloud computing have consequently become fundamental solutions to AI, enabling this kind of learning to be possible. Prior to 'training' and learning, computers were only capable of generating random results. After implementing learning mechanisms, computers can produce results that are logically coherent with the data source. Once this 'training' process has completed, it immediately develops into a new black box and learning is no longer required for future calculation, instead prior learning outcomes will be used and these outcomes will replace and become new 'training' material and data source, feeding the next cycle of iteration. This kind of learning begins very slow and gradually speeds up overtime at an exponential rate. Through learning, computers can efficiently find optimal solutions without needing to compute every data point but rather become proficient in narrowing down the magnitude of calculations, in such way achieving more precise conclusions in shorter time frames.