Artificial Intelligence (AI) has been a subject of headline news for a quite while. Google's AlphaGo's success in beating a human professional Go player on a full-sized board game created sensation about machine's rising capability. Similarly, IBM's Watson's success against legendry champions Brad Rutter and Ken Jennings, in 2011, winning the first prize of $1.0 million also made headline news. The granting of citizenship to humanoid has drawn attention of common people to increasing capability of human-like machines. Such development obviously has been creating the impression that machines are growing as a rival to human competency-with the apprehension that machines may take over human race. But are these machines posing risk of increasing inequality among countries, firms and individuals? There might be a debate whether machine will take over human race or not, but the exploitation of human-like intelligence capability appears to have been increasing inequality.
Often we are busy with technology updates having short-term effects. Although recent AI news may have little or no relevance to economic activities in near future, particularly in developing countries, their implications over next decade or so will likely be significant. In summarising Nobel laureate Paul M. Romer's work, it has been observed that it is easy to forget the long-run perspective of technology progression--the development of output, and more broadly human welfare, over decades or even centuries. Even small year-to-year differences in growth rates, driven by adoption of technology in product and production, which may seem tiny in a short-run perspective, cumulate. If such differences are systematic over decades, they build up to significant changes in competitive advantage, often leading to inequality between best performers and laggards. Long-run macroeconomic performance created by technology is thus a dominant driver of varying degrees of welfare enjoyed by individuals, firms and countries.
Technology play's a crucial role in market economies, in which actors develop new technologies through product and process-oriented research and development (R&D). Firms might develop, adopt and adapt new technologies in a bid to increasing the quality and reducing the cost for increasing profit. As a result, technology becomes an endogenous (internal) factor as opposed to weakly coupled exogenous (external cause) one. Despite the science fiction type impression of AI, a set of technologies forming the core of AI appears to have the potential of keep adding competence to firms in an incremental manner for enjoying the benefit of increasing the quality and reducing the cost of whatever they are producing over next couple decades. Moreover, such AI-based capability addition will keep decreasing the demand of skills in performing repetitive tasks, while creating the demand for creative, problem-solving capabilities. As a result, the rate of acquiring the capacity to benefit from AI will lead to increasing inequality between individuals and firms, consequently countries.
AI comprises five major technologies: computer vision, natural language, virtual assistants, robotic process automation, and advanced machine learning. Machine vision capability offers the opportunity of human like continuous visual monitoring of productive activities and product features alike. This technology has the capability to support innovations in detecting defects as soon as it happens leading to prompt action for attaining higher quality and reducing wastage. Moreover, robotic devices with the capability of human-like machine's vision can deal with variations with far higher accuracy, resulting in higher productivity and efficiency. Natural language capability coupled with machine learning leads to human-like capability on managerial and customer service jobs.
AI, as a family of technologies, has big potential to contribute to global economic activity. It's likely that companies will use these tools to varying degrees. For example, some will take an opportunistic approach, testing only one technology and piloting it in a specific function-while adopting all five in certain areas first and then absorbing them across their entire range of activities. Between these two poles will be many companies at different stages of adoption-resulting in varying degrees of competence development out of AI technology core. It's being estimated by researchers that AI could potentially deliver additional global economic output of around $13 trillion globally by 2030--about 16 per cent higher cumulative gross domestic product (GDP) compared with today. This amounts to about 1.2 per cent additional GDP growth per year. Such likely impact of AI on economic growth appears to be significant, when compared with previous major waves of technology-led transformation. For example, the introduction of steam engines during the 1800s boosted labour productivity by an estimated 0.3 per cent a year, while the impact from robots during the 1990s around 0.4 per cent, and the spread of IT during the 2000s 0.6 per cent. Well, such impact of AI may not be linear, but may build up at an accelerating pace over time, following a typical S-curve pattern. At the beginning, the benefit will be insignificant, but AI's contribution to growth may be three or more times higher by 2030 than it is over the next five years.
But all countries, firms and individuals will not equally benefit from this new growth driver. A key challenge is that adoption of AI could widen gaps between countries, companies, and workers. It's likely that AI leaders (mostly in developed countries) could increase their lead in AI adoption over developing countries. Leading countries could capture an additional 20 to 25 per cent in net economic benefits compared with today, while developing countries may capture only about 5.0 to 15 per cent. Countries such as Japan and China suffering from labour shortage due to the aging population will significantly benefit from AI.
At the firm level, AI technologies will likely lead to a performance gap between front-runners and slow and non-adopters. At one end of the spectrum, front-runners (companies that fully absorb AI tools across their enterprises over the next five to seven years) are likely to benefit disproportionately. It's being estimated by a lead think-tank that by 2030, they could potentially double their cash flow. On the other hand, laggards that do not adopt AI technologies at all or that have not fully absorbed them in their enterprises may experience by 2030 a 20 per cent decline in their cash flow compared with today's levels.
AI will likely widen the gap between individual's job prospects and earning levels. Job profiles characterised by repetitive tasks and activities that require low digital skills may experience the largest decline as a share of total employment, from some 40 per cent to near 30 per cent by 2030. The largest gain in share may be in non-repetitive activities and those that require high digital skills, rising from some 40 per cent to more than 50 per cent.
AI is no longer science fiction, nor a human-machine rivalry topic. It's about building human-like machine capability in making productive activities smarter, so that the quality goes up and cost comes down. Often such progress is made by delegating increasing roles from human to machines, and also by augmenting human capability. At the beginning, changes brought by AI may be insignificant, but in course of time the gap between leaders and laggards will accelerate, following a typical S-curve like pattern. On one hand, AI is a wonderful opportunity of producing more with less to meet our growing consumption, and on the other, it poses serious risk of widening the gap between individuals, firms and countries. It's time to respond to this opportunity in an appropriate manner so that all shares the new wealth likely to be created by AI.
M Rokonuzzaman Ph.D is academic and researcher on technology, innovation and policy.