In March 2020, the National Science Foundation (NSF) announced the first ten recipients of the NSF Program on Fairness in Artificial Intelligence in Collaboration with Amazon (FAI).
From the solicitation:
NSF and Amazon are partnering to jointly support computational research focused on fairness in AI, with the goal of contributing to trustworthy AI systems that are readily accepted and deployed to tackle grand challenges facing society. Specific topics of interest include, but are not limited to transparency, explainability, accountability, potential adverse biases and effects, mitigation strategies, algorithmic advances, fairness objectives, validation of fairness, and advances in broad accessibility and utility. Funded projects will enable broadened acceptance of AI systems, helping the U.S. further capitalize on the potential of AI technologies.
NSF Computer and Information Science and Engineering (CISE) Deputy Assistant Director Erwin Gianchandani was interviewed by Amazon Science about the challenges addressed by funded projects. He highlighted NSF’s public-private partnership with Amazon as a way to tackle some of the greatest challenges in Fairness in AI.
When asked about the “challenge of fairness in AI,” Gianchandani talked about “trying to get to an understanding of what fairness really means.” There are many definitions of fairness. The mathematical definition of fairness is “trying to ensure that the metric is consistent across [two] population[s] types.” If we can understand fairness perhaps we can “understand how to design our systems to build [it] in them.” Another challenge that Gianchandani mentioned was identifying who was responsible if an AI system makes an unfair decision. How do we think about accountability? We need to “provide the user with as much information as possible to minimize the likelihood of unfairness in the outcome-or at least provide an understanding of the types and levels of unfairness that may be inherent to the prediction from the AI system.”
The first ten funded projects address these challenges in four broad research areas:
- Ensuring fairness in algorithms and the systems that incorporate them — which begins with the definition and quantification of fairness;
- Accountability and transparency in AI algorithms;
- Using AI to promote equity in society; and
- Ensuring that the benefits of AI are available to everyone.
The value of the public-private partnership in addressing these challenges is “significant” as Gianchandani described.
“First, it’s valuable for our academic community to understand the kinds of challenges that industry is seeing… Beyond that, we all know that today’s AI revolution is grounded in large quantities of data that are readily available, along with compute resources to leverage those data sets… Third, academic researchers benefit from companies’ experience with accelerating the transition of research results out of the laboratory environment and into practice. Finally, another dimension that’s really important to us is training the next generation of researchers and practitioners…Providing our students who are studying fairness in AI with exposure to industry — to the problems that industry is facing — is a means to nurture the talent that our research ecosystem is going to need going forward.”
The latest solicitation for this award has a due date of July 13, 2020 and can be found here on the program site.
The Computing Community Consortium (CCC) has been working in the Fairness and AI activities space for some time. In 2018, CCC hosted a workshop on Fair Representations and Fair Interactive Learning. Then in 2019, CCC hosted the Economics and Fairness workshop, which produced this workshop report. Also in 2019, CCC initiated an effort to create a 20-Year Roadmap for Artificial Intelligence, which has an entire section on “AI Ethics and Policy.”
CCC also held a session at AAAS 2020 on New Approaches to Fairness in Automated Decision Making. See the summary blog to learn more about it.
