PAI Submits Response to OSTP, NSF Request for Information on National AI Research Resource

In accordance with the National Artificial Intelligence Initiative Act, the White House Office of Science and Technology Policy (OSTP) and the National Science Foundation (NSF) announced the formation of a National Artificial Intelligence Research Resource (NAIRR) Task Force earlier this year. This task force is charged with developing a roadmap for implementing NAIRR, intended as a shared AI infrastructure providing researchers and students with access to computational resources, data sets, educational tools, and user support. In July, the OSTP and NSF publicly requested comments to inform the work of the task force.

The Partnership on AI (PAI) was pleased to submit several examples of our projects and publicly available resources in response, highlighting our work around Demographic Data and Algorithmic Bias; Fairness, Transparency, and Accountability; Diversity, Equity, and Inclusion in AI; and inclusion and access to AI R&D through multistakeholder partnerships.

Please read our entire submission to OSTP and NSF here or an excerpt from it below:

Working with its multistakeholder community of partners in academia, civil society, industry and media, PAI is committed to increasing access, inclusion and participation in AI R&D.

One of PAI’s Partners, the Tech Policy Lab at the University of Washington, has extensive expertise in applying value-sensitive design approaches to technology policy. In 2019, PAI worked with the Tech Policy Lab to implement their Diverse Voices methodology within PAI’s ABOUT ML project. The aim was to solicit views and feedback from communities who are often the least likely to be consulted in the formation of machine learning system documentation practices that may impact them. The insights garnered through this consultation informed the inclusion of a glossary in the ABOUT ML resource library as well as the design and structure of the materials to promote clarity and navigational guidance to readers from diverse backgrounds.

Building from the lessons learned from the Diverse Voices team and the work of other responsible AI advocates, PAI launched the Methods for Inclusion research project that aims to enable AI researchers and developers to more effectively and ethically engage with a broad base of constituents and stakeholders in the development of their AI/ML projects. This work seeks to meaningfully include impacted communities in order to enable AI/ML developers to provide an array of products and services that can better meet the needs of diverse populations around the world, without further deepening existing social inequalities or generating harm. A forthcoming publication will identify a broad range of methodologies and practices that can be applied at different stages of the AI development process, drawing on the large body of scholarship that has grappled with the question of how to create inclusive channels of participation in other domains.

Drawing on these experiences at PAI, OSTP and NSF are encouraged to incorporate a focus on inclusion, participatory design, and democratizing access throughout all aspects of the development of the resource. Building a diverse community of stakeholders across sectors to engage and inform the development and deployment of the resource over time will be central to its success.