Forces Shaping AI Innovation and Applications:
Areas of Economic Impact
Summary
AI technologies are developed and deployed in complex environments. The confluence of all the market forces that inhibit and accelerate advanced technologies makes it difficult to forecast. One of the ways to handle this complexity is game theory. While it is often tricky to capture all the complexities, and in this case, it certainly is, the general idea seems intuitively correct: the economic process will settle at an equilibrium point that satisfies the external conditions. Additionally, factors defining the environment are dynamic in a short time and hard to predict long-term trends, which undoubtedly contributes to a lot of uncertainty and movement of the equilibrium. This brief introduction is part of a more extensive study that attempts to find where that equilibrium point might be: somewhere between the regulatory forces, the court of public opinion, the scientific curiosity of researchers, and powerful forces that drive economic growth.
1.Economic factors and innovation
Economists identify many factors that drive economic growth or contribute to the reversal of trends into negative territory, resulting in economic stagnation. Many of those factors are closely correlated, such as human and physical capital or the availability of natural resources.
Source: https://www.sciencedirect.com/science/article/pii/S2405844021012123
In addition, the political environment plays a significant role and can be correlated with trade, level of innovation, and capital.
A confluence of the above factors drives a more aggregate driver like productivity, which describes a relationship between inputs that go into economic processes vs. the obtained output. The productivity rate increases if we produce more with fewer inputs (capital, energy, time, etc.) As a result, on the macro-scale, increased profits can be reinvested and generate positive feedback for further capital growth, more people participating in innovation, increased consumption of the economic output, etc. This simplistic model outlines very few dependencies that influence health and economic prosperity, but all these factors are places where advanced computing technologies, including AI, play a significant role.
Source: https://link.springer.com/article/10.1007/s40821-020-00172-8
In this context, we can see significant incentives in the economy to drive growth and build capital. But, unfortunately, those driving forces may not prioritize responsible design and usage of AI. Moreover, advanced information processing technologies might be seen as one of the critical catalysts for further growth, much stronger than advances in traditional areas like engineering and science, e.g., mechanical engineering or material science. Moreover, some publications suggest that positive innovation effects driven by rapid progress in computing supersede the adverse effects of stagnation in older sciences.
Source: https://www.sciencedirect.com/science/article/abs/pii/S0094114X17315410
2.Innovation and Productivity
The new information processing and computing technologies hold great promise. They act as a catalyst for many capital-intensive industries. For example, energy production, agriculture, or trans-oceanic shipping can save billions of dollars by optimizing their processes by a fraction of a percent. Computing creates many unique and often unexpected opportunities that translate into economic growth and disruption of traditional business processes. Increased productivity might not be the only effect. Sometimes, entire new industries get created, replacing obsolete ones. Sometimes, that change might have a positive social impact, but it does not always.
Source: https://libres.uncg.edu/ir/uncg/f/A_Link_Basic_1981.pdf
Once we explore data in specific industries, the numbers paint a mixed picture. For example, productivity is often cited as a critical indicator of the economic condition on a macro scale. Productivity and economic impact are still rising in information-intensive areas that rely on advanced algorithms. However, there are other areas of the economy where computing is far less prevalent as a central business driving force. Those more traditional industries that rely on manual labor and older technologies will be under increasing economic pressure to modernize tooling, value creation processes, optimize supply chains, and energy consumption.
Source: https://www.sciencedirect.com/science/article/abs/pii/S0954349X1200063X
Automating economic processes with technology is a complex problem that impacts society. Automating blue-collar and white-collar jobs produces side effects in the labor market. It might hurt wages and systemically displace many employees. Shifting labor markets leads to changes in regional economic activity. Educational institutions may not be able to react quickly to market changes resulting in demand for high-skilled workers outpacing the supply.
3.Jobs and Social Implications
Economists often quantify economic processes in very general terms. Productivity or innovation levels are macro-level factors correlated with national GDPs or global economic conditions. More subtle effects are in play on the micro-scale and may be precursors for future large-scale consequences. For example, narrow specialization for many technology workers requires years of training. As a result, the job market for specialists becomes more shallow as the technologies continue to spread, new tools appear on the market, and existing technologies quickly become obsolete. Each new invention contributes to innovation on the macro scale, which is an economic positive. On the other hand, business disruption and increased uncertainty may negatively affect the job markets. For example, fewer students and high-tech workers might see a successful career path in a rapidly changing technical landscape.
Source: https://link.springer.com/article/10.1007/s10290-009-0009-2
Our dependence on computing is more prevalent, a risk that society must consider. Will we permanently lose specific skill sets that are successfully automated or replaced? Human knowledge, as a concept, is difficult to quantify until a problem forces us to act. The more advanced the technology, the higher the impact. What will be the cost of replacing creative jobs that require years of training? There is a potential risk of losing control over critical systems that evolved beyond our comprehension, which we limit by our innovation.
Source: https://www.sciencedirect.com/science/article/abs/pii/S0263237399000195
4.Definition of "responsible" AI
The economic and regulatory environment is quite dynamic. Therefore, compiling and synthesizing all global initiatives affecting multinational organizations is quite a task. We must build models that evaluate environmental sustainability, privacy, and social benefits and costs. The list of factors is constantly growing.
Source: https://www.sciencedirect.com/science/article/abs/pii/001429219400075B
A common denominator for all initiatives must include transparency and accountability for designers, developers, and users. In addition, there is a growing awareness that legal responsibility and difficult insurance questions might soon be driving the regulatory initiatives. The environmental sustainability of AI computations will also be in the spotlight. Blockchain and AI might quickly become the highest energy consumers on our planet. Many believe that there are enormous challenges if we want to continually innovate with enormously costly neural network models that need megawatts of power to train and operate.
The definition of responsible AI will continue to be redefined in scientific circles, legal communities, and legislative bodies.
