Federal Advisory Group Report Focuses On Diversifying AI Research and Development
Yuichiro Chino/Getty
The National Artificial Intelligence Research Resource Task Force issued a report outlining a framework for national AI research and development, hinting on equitable datasets and computing access.
The National Artificial Intelligence Research Resource Task Force released a fresh report on Wednesday that lays out a plan to develop a standardized cyberinfrastructure across the artificial intelligence research field in the hopes of fostering responsible AI applications.
Equitable access to computational research data is a primary feature of the group’s report. Co-chairs of NAIRR discussed how researchers affiliated with wealthy private companies or universities with steady endowments have disproportionate access to the computational power and data to conduct advanced research.
“We must expand the ability of America's research community to pursue cutting edge AI research and make the pathways to do so more equitable,” Lynn Parker, the NAIRR co-chair, said during a press call. “The NAIRR seeks to lower the barriers to entry to allow more Americans to get involved in AI research and development to prevent unintended consequences or disparate impacts from the use of AI.”
The interim report––the first of a two-report series––is composed of recommendations for federal agencies to reference when studying new AI technologies. Using the NAIRR’s proprietary user access portals, researchers would be able to have access to a diverse amount of computing resources, as well as connect users to the cloud and high performance computing servers.
Manish Parashar, the fellow NAIRR co-chair, told reporters that the user portals would also contain datasets to give researchers access to a “robust, transparent, and responsible data ecosystem.”
“The NAIRR would also act as a hub for AI testbed and test datasets, cataloging, and providing access to existing testing and benchmarking resources that can accelerate progress in AI R&D [research and development] and increase access to the research community to these high quality test beds,” he said.
Given the large volume of potentially sensitive data stored across the new network, the task force also advocated the use of a zero trust standardardized architecture to safeguard databases.
While NAIRR leadership is keen on creating incubator resources to expedite the development of new AI technologies, the report also emphasizes best practices for responsible and ethical AI.
“As a nation, we must come together to expand access to the resources that fuel AI, providing pathways for more Americans to pursue AI research and development (R&D) and access state-of-the-art resources,” the report’s executive summary reads.
These pathways to removing civil liberty threats within AI tech include better testing procedures, bias mitigation, and improved safety features. Incorporating diverse perspectives and backgrounds into the development of AI applications is also a priority within NAIRR’s report. Parker noted that keeping datasets and computing power accessible will help make AI development widely available.
“The strategic objective we're establishing in NAIRR would be to strengthen and democratize the US AI innovation ecosystem in a way that protects privacy, civil rights and civil liberties,” she said. “AI is transforming our world and a growing resource divide between those who have access to the resources needed to pursue cutting edge AI, and those who don't, threatens our nation's ability to cultivate an AI research community and workforce that reflects America's rich diversity.”