NSF Institutes Develop AI for Public Benefit

With AI affecting more aspects of our lives and the world around us, all statisticians and data scientists need to develop and apply emerging technology to benefit everyone. To meet this challenge, the National Science Foundation established 25 national artificial intelligence research institutes with the mission of advancing a cohesive approach to AI-related opportunities and risks. A key aspect of the AI institutes is to consider the ethical challenges presented by new technology, including fairness, transparency, and security. In addition to being the center for cutting-edge scientific research, the AI institutes are driving the development and implementation of AI to serve the greater good.

The institutes are supported by the National Science Foundation, but some have partnered with organizations such as the National Institute of Standards and Technology. The institutes focus on six broad areas of AI research, starting with basic research on the foundations of machine learning. Areas of research include the statistical basis of machine learning algorithms, intelligent networks and edge computing, and the
interaction between humans and AI.

The institutes are developing the basic technology underlying all AI. Many leading universities are participating in this research to drive AI capabilities forward, including The University of Texas, Columbia University, Georgia Institute of Technology, The Ohio State University, Washington State University, and Duke University.

One of the most important areas of interest in AI—and sometimes an area of concern—is the impact on and use of AI in education. Several institutes concentrate on developing applications and tools to leverage AI in education for maximum impact while promoting ethical best practices. Others focus on the educational process itself, including engaged learning at North Carolina State and student-AI collaboration at the University of Colorado Boulder. Still other institutes focus on education for specific groups of people, such as adult learning and online education at Georgia Tech, workforce development at Washington State University, and children with speech and language challenges at the University at Buffalo. The AI Institute for Inclusive Intelligent Technologies for Education at the University of Illinois leverages AI to support all K–12 students with the goal of achieving education for everyone in a diverse, inclusive, and highly effective manner.

Almost half the institutes focus on developing AI applications in a particular area, leveraging the technology to support key areas that are important for economic growth. Some target intelligent applications in natural sciences applied to manufacturing. Another group develops intelligent applications in cyber infrastructure, security, and threat analysis. Several institutes focus on agriculture, with a particular interest in resilient agriculture and the effects of climate change.

The institutes have developed important partnerships with organizations and agencies. For example, the Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography at the University of Oklahoma works closely with NASA scientists and AI experts as a leading member of the new Science Understanding Through Data Science consortium, which will be featured in a future Stats4Good column.

From the perspective of Data for Good, the AI institutes perform high-impact work. Two stand out as leaders in leveraging AI for society’s general benefit. One is the AI Institute for Societal Decision-Making at Carnegie Mellon University, which brings together social scientists and AI experts to develop solutions to better understand and improve human decisions using AI. The AI Institute for Trustworthy AI in Law and Society explores the question of trust in AI. This collaboration between four universities—the University of Maryland, Georgia State University, Morgan State University, and Cornell University—seeks to understand how AI is perceived, determine what constitutes trust in AI and how to promote it, and develop policy needed to foster trusted and trustworthy AI.

The AI institutes established with the National Science Foundation’s support are leveraging AI to create powerful avenues of change where statistical science is the key catalyst in a technological revolution. The growing impact of the AI institutes is a call to action for the statistical scientific community. Statisticians and data scientists in all areas and disciplines are needed to support these initiatives in developing effective and trustworthy AI for public benefit. In addition to direct participation of these institutes, leveraging the publicly available data and resources they develop will have a huge effect on programs that create a better future for everyone through Data for Good.

David Corliss

With a PhD in statistical astrophysics, David Corliss works as a data scientist in industry. He serves on the ASA Board as a Council of Chapters representative and is the founder of Peace-Work, a data for good nongovernmental organization.