AI R&D: Bridging the Resource-Access Divide

From serving up an unexpectedly perfect song in your music app to charting the safest evacuation routes for residents trapped by wildfires, AI is having an increasingly pervasive and impactful role in human life.

In the United States, however, the massive amounts of power and data AI research and development currently requires limits this work to a relatively small subset of researchers.

This AI resource divide not only risks perpetuating bias in AI models, but it also prohibits the emergence of an AI research com­munity and workforce that reflect the country’s rich diversity and ability to harness AI for the public good.

From June 2021 through April 2023, a U.S. advisory group developed a detailed plan to strength­en and democratize U.S. AI innovation by creating the National AI Research Resource (NAIRR) cyberinfrastructure for AI R&D.

In a recent Computer magazine article, “Strengthening and Democratizing Artificial Intelligence Research and Development,” NAIRR Task Force researchers report on this crucial work.

The Vision: An Accessible AI Cyberinfrastructure

To democratize AI efforts and sustain U.S. competitiveness over the long term, the NAIRR Task Force proposed the creation of a widely accessible cyberinfrastruc­ture that will empower diverse users across fields by giving them access to computational, data, soft­ware, and training resources and testbeds.

The NAIRR infrastructure will be created by leveraging, linking, and augmenting existing cyberinfrastructure resources. Among its benefits, the NAIRR will

• Allow cutting-edge explorations and encourage collaboration across dis­ciplines
• Cre­ate opportunities to train the future AI workforce
• Support and advance trustworthy, responsible AI
• Catalyze the development of ideas that can be practically deployed for societal and economic benefits

As the article notes, this federated research resource will “create pathways to broaden the range of researchers in­volved in AI and to grow and diversify approaches to, and applications of, AI.”

The NAIRR’s ability to create opportunities for prog­ress across scientific disci­plines will include critical areas such as AI auditing, testing and evaluation, trustworthy AI, bias mitigation, and AI safety.

NAIRR’s Four Goals

The task force set four measurable goals for the NAIRR:

• Spur innovation
• Increase diversity of talent
• Improve capacity
• Advance trustworthy AI

To meet these goals, the NAIRR will support researchers and students from diverse backgrounds who are pursuing foundational, use-inspired, and translational AI research. These users will be U.S.-based and/or affiliated with U.S. orga­nizations, including academic insti­tutions, nonprofit organizations, and startups or small businesses.

NAIRR Administration and Governance

After considering multiple models, the task force consensus was for a cooperative stewardship model for NAIRR administration and governance as follows:

• A single agency—the National Science Foundation—will oversee NAIRR and serve as its administrative home.
• An in­teragency steering committee will be active in NAIRR decision-making.
• An external nongovernmental operating entity will run the day-to-day NAIRR operations.

Because the NAIRR will require expert advice on everything from tech­nical resource design to user experience and regulatory compliance, the task force also recommended the establishment of four advisory boards:

• The Science Advisory Board will focus on changing needs across scientific domains to facilitate rapid innovation support.
• The Technology Advisory Board will focus on cutting-edge solutions related to technology infrastructures, workforce training, and privacy and security.
• The Ethics Advisory Board will focus on issues related to ethics, fairness, bias, and accessibility, as well as advising on blind spots and emerging risks.
• The User Advisory Board will focus on how NAIRR can serve its users, advising on the new directions and emerging user perspectives.

Dig Deeper

AI is an essential engine of innovation that is driving scientific discovery and eco­nomic growth. It is also increasingly emerging as an integral component in solutions that impact everything from our routine daily tasks to changes across societies.

As the NAIRR Task Force points out, this immense power is coupled with a potential for negative social, environ­mental, and economic consequences. To realize AI’s positive, transformative potential, it must advance in a way that protects privacy, civil rights, and civil liberties, while also pro­moting fairness, accountabil­ity, transparency, and equity.

These opportunities, along with further details on the task force’s vision and plan for NAIRR, are discussed in depth in “Strengthening and Democratizing Artificial Intelligence Research and Development.”

To dig even deeper into these and other issues, join other AI experts, researchers, government officials, and enthusiasts at the international IEEE Conference on Artificial Intelligence (IEEE CAI) 5–7 May 2025 in Santa Clara, California.

In addition to showcasing the latest AI research and breakthroughs, IEEE CAI emphasizes applications and key subject areas, from sustainability and human-centered AI to issues and industry-specific applications in healthcare, transportation, and engineering and manufacturing.