Preparing for the age of AI machines

The memory of a series of articles published in The Economist on "Autonomous Vehicle" during the first week of March 2018 is as fresh as morning dew. To welcome these smart machines, China has announced its plan to build an 'intelligent superhighway' for self-driving vehicles at a cost of more than $2 billion, as reported by the media. Bloomberg has reported that South Korea is planning $64 billion in smart highways. On the other hand, to protect the jobs of drivers, India has opted for creating a barrier to the entry of autonomous vehicles. Although the United States of America (USA) intends to accelerate the pace of deployment of such vehicles on America's highways, bipartisan legislation to promote safe development of autonomous vehicle technology remains stalled in the US Senate over safety concerns. There are also worries about data privacy and cyber security as machines evolve into smart devices that can access user's information. Often the risk of hacking of autonomous machines to cause deliberate harm is also a serious concern. In addition to preparing the infrastructure, improving the safety, dealing with cyber security and being concerned with jobs, what are the other transformations we should prepare for?
Exciting visuals, hype-fuelled media reporting, and Hollywood movies have created significant excitement and also apprehensions about artificial intelligence. There is no denying that the entry of new technologies and innovative products often creates such enthusiasm or even anxiety. Often varying interpretation of artificial or machine intelligence creates perplexity. Intelligence is a suitcase word, having many connotations. Often Intelligence has been defined as one's capacity for logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, and problem-solving. If machines have these capabilities, what is wrong with it? Does it mean that machine will grow as an adversary of human beings? Will they conspire against humans?
If we think so, why are we developing them? In order to be an adversary, or acquiring the capability to conspire against the human race, machines should have the capability to imagine to come up with ideas unknown to opponents? So far, artificially intelligent (AI) machines are built to automate the execution of doing jobs, which human beings have mastered? Even in doing so, there has been a tremendous difficulty. For example, it's taking decades of research to build machine capability to automate the driving skills. Moreover, there has been no indication that we can develop machines to imagine. If the reality in building AI machine is not so sexy, why is it a big deal?
Basically, machines are being developed to delegate roles from humans to technology to get the jobs done better at less cost. Our success of delegating roles of manipulation and supplying energy to machines increased our productivity and empowered low-skilled people to contribute to replication of industrial products. As a result, income grew and increasing number of jobs were created on factory floors for low-skilled people to replicate the same products, often millions of times. The growth of communication technology and computer opened the opportunity of distributed production facilities, creating jobs across the world. Moreover, dangerous jobs were delegated to machines making the workplace safer than ever before. As a matter of fact, the progression of building better machines came as a blessing for the human race. Then should we be concerned about the continued progression of machines?
Until 2000, despite the growth of the capability of machines, human beings were indispensable in production, as machines could not sense, and perceive the situation to decide about actions on their own. Even two decades ago, technology was far behind than human eyes to generate high-resolution images. But due to the phenomenal growth of digital image sensors, particularly fuelled by smartphone revolution, low-cost micro digital sensors can produce very high-resolution images. In addition to image sensors, many other sensors are being developed to sense diverse variables starting from smell to chemical content, or hidden defects of product. For example, thermal imaging sensors sense the presence of hidden air bubble in clay pots, or ultrasonic sensors extract data about the internal ripeness of watermelon. Progress is being made in processing those data to perceive the situation to decide whether the product is ready to be moved to the next phase for processing or to determine the grade of the food. Such progression is reaching a state to build machines having the human-like capability, even enhanced in certain cases, to sense and perceive the situation to take smart actions on their own. As a result, human-free production is getting momentum in an accelerated space. The age-old plain simple economics is driving this journey: produce better products at lower cost to increase profit.
Human-free production possibility is giving rise to concerns for jobs. How will the wealth be distributed in an equitable manner? Will the common people lose the confidence in their ability to give the effort to earn? Such questions pose serious concerns. Moreover, human-free production is also increasing the monopolistic market power. If companies of the Western countries can produce better products at a lower cost without the labour of developing counties, how can developing countries get a share of the industrial economy is a grave concern. By capitalising on rapid technology progression, if high performing companies keep acquiring the capability of setting the price to make a profit, while compelling competitors to take a lower price to incur a loss, the market runs the risk of being monopolised.
Such possibility creates the threat of human-free monopoly-making both citizens and the governments across the globe vulnerable. Does it mean that the underlying capability of open market economic theory will keep experiencing erosion in its potential to govern production, distribution, and consumption of wealth? It appears that the continued progression of artificially intelligent production technology is creating confrontation between economics, governance and social order. But in the absence of continued progress in building smarter machines, we cannot deal with the challenge of producing more wealth from depleting resources, while causing less harm to the environment, to meet growing consumption in a sustainable manner.
Such a reality for wealth creation within given constraints and the role of artificially intelligent machines, posing threat to core underpinnings of governance, has created the demand for broader transformation to prepare for the age of artificially intelligent machines.
M Rokonuzzaman Ph.D is academic, researcher and activist on technology, innovation and policy. [email protected]