Registration
You will receive an email confirming your registration.
Given the rapid advancement of Artificial Intelligence (A.I.) and its associated effects, there is a need to model efforts that allow us to explore A.I.’s development and its societal impacts.
During the Global Technology Summit 2018, Carnegie India hosted David K. Bohl, senior research associate at the University of Denver’s Frederick S. Pardee Center for International Futures, for a student workshop that explored how artificial intelligence will shape economic and human development over the coming decades. This workshop was part of a new initiative called KnowledgeTransfer@CarnegieIndia, which aims to provide a platform for genuine exchange of ideas and knowledge among students, practitioners, experts, and other interested audiences.
DISCUSSION HIGHLIGHTS
- International Futures Model: Participants discussed the International Futures (IFs) model, developed by the University of Denver, which is a technology-enabled, long-term, global, integrated forecasting platform. It merges forecasts across different sub-models—such as population, economy, and energy, among others—¬and the interconnections allow IFs to simulate how changes in one system may or may not lead to changes across all other systems, they said. Participants cautioned that this model allows one to extrapolate the future, not predict it. To better understand the model, they discussed how the model forecasts India’s population over the next decade, based on drivers including the birth rate, death rate, and the net migration rate. They also noted that each of these drivers are in turn influenced by others. For instance, the rate of communicable diseases, among other factors, impacts the death rate. Barring any stark aberrations in these drivers during the period under consideration, the model extrapolates trends to provide a reasonably reliable idea of India’s population 10 years in the future, they said. Participants discussed that the model helps decision makers across governments, international organizations, think tanks, non-governmental organizations, and the private sector to create models that evaluate possible outcomes of various courses of action.
- Technology and Policy: Participants observed that technology has gone from being a tool at our disposal to being an integral part of our environment and a way of life. Participants noted the disruptive potential of A.I. and how it is transforming individual experiences, the nature of work, societies, and their governance. They discussed that A.I.’s rapid adoption necessitates the adoption of rules and regulations that govern its use. For this, it is important to understand A.I. and its impact at the national, regional, and global levels, they said. Participants discussed that the IFs model will be expanded to model and forecast A.I. and its impact.
- Modeling A.I.: Participants discussed that creating a model that is complex enough to gauge the effects of the adoption of A.I., requires careful consideration of the technology’s development and associated impacts. First, one must understand the factors driving the development of A.I., including hardware and software development, and establish historical benchmarks for measuring progress in A.I.’s development and impact, they said. To achieve this, participants discussed the importance of examining historical trends, considering literature on the subject, and involving subject matter experts. Second, participants pointed out the importance of considering how A.I. will impact various sub-models, including socio-political systems, businesses, work force, and heath, among others. They observed that this requires careful consideration of A.I.’s development and associated impacts on various economic, political, social, and geopolitical systems, including health, productivity, and inequality, among others.
- Limitations of IFs: Participants discussed the limits of using models to forecast the future. First, they discussed that data sourced from international organizations and national governments are sometimes incomplete or inaccurate. Second, while accurate data can be modelled to extrapolate future trends, they cannot predict sudden aberrations such as the 1973 oil crisis, they said. They also noted that while these may appear to be a huge distortion in the short term, they are usually just slight variations from the normal course of events in the long term. Third, the accuracy of forecasts also depends on whether they are based on data related to outcomes or expenditures.
This summary was prepared by Arjun Kang Joseph, an intern at Carnegie India.