NSF News

New NSF-Australia awards will tackle responsible and ethical artificial intelligence


The U.S. National Science Foundation, in partnership with Australia’s national science agency, CSIRO, is announcing grants totaling $1.8 million on the U.S. side and $2.3 million on the Australian side to accelerate groundbreaking research in responsible and ethical artificial intelligence solutions to societal challenges, including pandemic preparedness, drought resilience and harmful environmental emissions.

Responsible and ethical AI concerns have grown sharply with the increasing availability of AI-powered technologies. Awards under the NSF-CSIRO partnership are expected to contribute to establishing ethical frameworks — and ultimately guidelines — to ensure AI algorithms and their deployments are safe, fair, and beneficial to all citizens.

"NSF's continuous efforts to form collaborations with nations across the world represent part of our core mission of ensuring that the research and discoveries that result from our investments benefit citizens everywhere and address the priorities and challenges of our like-minded partners," said NSF Director Sethuraman Panchanathan. "I want to congratulate the U.S. and Australian teams on their recent awards. There is much to be done in the field of responsible and fair artificial intelligence, and we are eager to see how their research accelerates and innovates solutions that help solve critical challenges across AI-powered technologies."

"CSIRO is proud to stand side-by-side with the NSF, committed to responsible and ethical AI not only through our own research, but by catalyzing new opportunities across the ecosystem," said CSIRO Chief Executive Larry Marshall.

Proposals submitted by the teams of U.S. and Australian researchers in response to the NSF Dear Colleague Letter addressed the design, data use and algorithmic aspects of AI systems as they relate to several societal problems. For this round of awards, the proposals addressed the following topical themes:

  1. Drought resilience, toward net zero, infectious disease resilience

    Fair Sequential Collective Decision-Making

    Led by researchers at the University of Nebraska-Lincoln and Rensselaer Polytechnic Institute on the U.S. side, this project aims to develop AI-powered approaches that enable responsible, fair, and equitable solutions to drought, environmentally harmful emissions, and infectious disease. AI will be employed to determine equitable allocation of resources such as water, optimal placement of refueling stations for non-fossil fuel vehicles, and vaccines and other medical supplies.

    The U.S. team will be joined by Australian researchers at the University of New South Wales.

  1. Infectious disease resilience

    Understanding Bias in AI Models for the Prediction of Infectious Disease Spread

    Led by researchers at Emory University, Arizona State University and George Mason University on the U.S. side, this project aims to mitigate bias in AI-powered modeling and prediction of disease spread for pandemic prevention and response. To accomplish this objective, the teams will investigate how biased data spreads to modeling pipelines and leads to biased AI solutions. In addition, the researchers will leverage different metrics of fairness in AI and study how these fairness measures can be incorporated into AI optimization procedures to mitigate bias.

    The U.S. teams will be joined by Australian researchers at the University of New South Wales and RMIT University.

    Graph Representation Learning for Fair Teaming in Crisis Response

    Led by researchers at UCLA and The University of Texas at Austin on the U.S. side, this research aims to understand how scientific communities have responded to historical pandemic crises and how to improve the response to future pandemics through fair teaming solutions. Researchers in this project will develop fair AI models for establishing global scientific communities that provide equal and inclusive working environment.

    The U.S. teams will be joined by Australian researchers at the University of Technology Sydney and the University of Melbourne.