OSTP and NSF launched the task force last Thursday. Under the 2020 National AI Initiative Act, the task force will look at how to expand access to AI education and other critical resources. The task force includes members from NIST, the Energy Department and top universities.
Lynne Parker, the director of OSTP’s National AI Initiative Office, will co-chair the task force along with Erwin Gianchandani, the NSF’s deputy assistant director for computer and information science and engineering.
Parker previously served as the deputy U.S. chief technology officer under the Trump administration.
As part of this rollout, OSTP will also create a National AI Advisory Committee, which will provide recommendations on topics that include AI ethics, research and development, and AI’s impact on the workforce.
The task force will submit two reports to Congress — an interim report in May 2022 and a final report in November 2022.
OSTP Director Eric Lander, who also serves as President Joe Biden’s science adviser, applauded the “foundational investment” in technology leadership.
“The National AI Research Resource will expand access to the resources and tools that fuel AI research and development, opening opportunities for bright minds from across America to pursue the next breakthroughs in science and technology,” Lander said.
NSF Director Sethuraman Panchanathan said the task force will have an essential role in driving new breakthroughs in AI.
“By bringing together the nation’s foremost experts from academia, industry, and government, we will be able to chart an exciting and compelling path forward, ensuring long-term U.S. competitiveness in all fields of science and engineering and all sectors of our economy,” Panchanathan said.
NSCAI sees common DoD ‘digital ecosystem’ for AI development
Meanwhile, the Defense Department is moving ahead with its own AI rollout, guided in part by recommendations drafted by National Security Commission on AI in its final report.
David Kumashiro, NSCAI’s director for research and analysis, said the commission’s 750-page report generally reflects on the need for DoD and the federal government to catch up with the private-sector investments in AI.
“AI is going to be ubiquitous across all aspects of military affairs, so I think the right question is, what areas is AI not going to touch? I would challenge folks to really list out, in those terms, where we don’t think AI is going to be a critical component,” he said last Thursday in a virtual panel hosted by the Center for Autonomy and AI.
The final report urges DoD and the intelligence community to make foundational investments in order to be AI-ready by 2025. That work includes giving DoD personnel greater fluency in digital tools.
“AI-ready by 2025 is really just about your baseline digital literacy and access to the tech stack, and software and data that allows you to integrate,” Kumashiro said.
To bridge the gap between the warfighter and technologists, the commission recommends standing up AI delivery teams within each combatant command. The Army, Kumashiro added, has made progress by standing up tactical data teams, but added DoD should take a DevOps approach to AI development.
“We really should shift this mindset that the technology is going to have this roadmap, milestone chart like traditional acquisition programs, and instead it will be much more much more iterative and require much more involvement with the warfighter to refine where this technology is going,” he said.
Success also depends on IT infrastructure that enables collaboration. DoD’s Joint AI Center stood up its Joint Common Foundation in March to accelerate testing and adoption of AI tools across the department.
DoD also has stood up platforms such as Platform One for enterprise DevSecOps services, and the Navy supports Black Pearl — a group of military, civilian and contractor personnel with experience with software delivery challenges.
“We need to start linking them together, we need to start identifying what are those best tech stacks that are out there, stop reinventing that, and really start committing to a few of them. I’m not saying ‘Hey, we’ve all decided on this one.’ I think as long as you’re keeping with good commercial standards and practices of containerized data management, it shouldn’t be a problem of linking and networking all of these different platforms into this larger digital ecosystem,” Kumashiro said.