Either innovate or keep getting poorer

Advances in artificial intelligence (AI) transforming nearly every sector, from transportation to medicine to defence, have been dramatic. So it is worth considering what will happen when it becomes even more advanced than what it already is. Top-level computer scientists are failing to reach a consensus on the likely consequential effects of AI. This lack of convergence of understanding has a profound implication - policy makers do not get guidance to leverage beneficial aspects of AI, while minimising its negative effects.

There is no denial that there have been many false starts of AI. Despite many science fictions and movies causing fear in people's mind about human-like machines, the negative implications on human life, particularly on jobs, have remained under-stated.

Honda's ASIMO looks and behaves very much similar to human, but so far it has not taken jobs; rather it has entertained human beings. On the other hand, some of the AI machines do not have conventional outfits similar to human, but they invisibly compete with humans to get the job done better, cheaper and safer. For example, autonomous cars do not require human drivers, nor do they have human-like robot drivers in the driving seats. It is presumed that with the maturity of the self-driving smart cars, there will be fewer accidents, making roads safer. Moreover, due to being more responsive than human drivers, autonomous cars will be driving faster requiring less inter-vehicle distance, consequentially increasing throughput of existing highways.

Whether such uprising of machines is good or bad sparked furious arguments two centuries ago as industrialisation took hold in Britain. Over the last centuries, it has been found that machines have become blessing for the growth of human civilisation. What will happen, this time?

What are likely implications? The Economist has reported that according to the study of Carl Benedikt Frey and Michael Osborne of Oxford University, published in 2013, 47 per cent of jobs in America were at high risk of being "substituted by computer capital" soon.  By 2025 the "annual creative disruption impact" from AI could amount to $14 trillion-33 trillion, including a $9.0 trillion reduction in employment costs, thanks to AI-enabled automation of knowledge work, predicts Bank of America Merrill Lynch. Cost reductions of $8.0 trillion in manufacturing and health care; and $2.0 trillion in efficiency gains from the deployment of self-driving cars and drones, the prediction adds.

What is the stake of developing countries in this round of uprising of machines? In order to imagine likely implications on developing economies, we need to go to the basic roles of humans and machines in production. Human beings have seven major roles to play in wealth creation: 1. Idea generation, 2. Sensing, 3. Perception and memorisation, 4. Planning, reasoning and deciding about actions,  5. Coordination and communication, 6. Manipulation, and  7. Locomotion and energy.

With the growth of steam engine, the role of locomotion and providing energy was taken over by machines. The development of mechanical tools, including human hand-like robotic hand, started taking over the role of manipulation. The telecommunication, business software and computers are taking over the role of coordination and communication from human. The ability to analyse data to figure out inter-dependence of tasks and capabilities of different actors is rapidly growing to enable machines, primarily software, to take over the planning role from human as well. The pattern recognition and statistical methods of data analysis, whether produced by real-time or off-line sensors, and deep mapping between variables/actors is enabling machines to perceive the situation. Due to the growth of digital imaging, primarily fuelled by smartphones, machines can now have visual sensors, with far higher resolution than human eyes, once thought to be an impossible target to meet. Moreover, due to the rapid reduction of cost and size, and improvement of other powerful sensors, like LIDAR or RADAR, it's increasingly becoming realistic to build machines having far better sensors than human beings have.  Therefore, the 2nd most important capability of human beings is also rapidly getting substituted. The only one left is: the idea generation capability. So far, there is no clue, whether it's possible to replicate this capability in machines.

Cost of AI machines is getting cheaper. There is no doubt that the possibility of developing machines having the capability to replace basically all roles of human in production is getting very real. Another important question to ask is: whether the cost of developing capability of such machines to replace human roles is a profitable option. The reality is that the cost of AI machine is rapidly falling. The growth of low-cost, high-power computing and sensing capability, primarily fuelled by PC and mobile phone revolution, is the underlying driver of such rapid reduction of cost, making them cheaper than human labour.

Reversal of globalisation has begun. In the 1960s, production moved from advanced countries to developing countries to take the advantage from low-cost human labour. It was termed globalisation. The early entrants--like Malaysia, Thailand and China--benefited from such globalisation. Now replacing human roles with machine is getting cheaper due to rapid reduction of cost for increasing capabilities of machines.

The decision of Adidas of moving its plant from China back to Germany appears to be the beginning of reversal of manufacturing job movement from the advanced countries to the developing ones. The race of having smarter machines to produce better quality product at less cost than human beings appears to have no end. The possibility of increasing income level by creating the opportunity to deploy low-cost labour force to produce for export is basically eroding. It appears that machines are going to put ceiling on labour wage, which will be continuously drifting downward.

For developing countries, the scope of sharing global wealth by supplying labour in manufacturing appears to have started shrinking. This trend will simply keep progressing creating downward pressure on labour wages, making the investors and smart machine innovators richer than ever before.

Either innovate or keep getting poorer is a reality to face. In this scenario, developing countries must not solely rely on labour supplier-based development model. They must invest in innovating production processes to deliver better quality outputs at lower cost. Such capability will not only enable them to sustain competitiveness of their production process, but it will also open the acquisition of the capability of being smart machine innovators. Being smart machine innovators, they will also get growing share of new wealth of the world. Otherwise, the growth of AI smart machine will be continuously eroding wages of the labour, making developing countries poorer. So getting into building AI machines is no longer an option to enter into innovation economy; it's rather going to be the only option to survive and support the real income growth of citizens.  To deal with such emerging situations, along with developing economic zones, developing countries must also focus on innovation universities and, most importantly, policy reform. 

Rokonuzzaman, Ph.D is an Academic, Researcher and Activist: Technology, Innovation and Policy.

[email protected]