WASHINGTON: The The Pentagon has released its long-awaited Responsible Artificial Intelligence (RAI) strategy and implementation path, recognizing that the Department of Defense will not be able to maintain a competitive advantage without transforming itself into an AI-enabled and data-centric organization, the RAI as a salient feature.
The company-wide strategy signed by Deputy Secretary of Defense Kathleen Hicks and released Wednesday [PDF]prepares the Department of Defense for the next step in its AI journey Outlined several points of action around testing and assessment requirements and strengthening its digital workforce.
“The implementation of RAI in the Ministry of Defense will not succeed with a set of rigid, uniform requirements,” the strategy says. “A flexible approach is needed to encourage innovative thinking, as requirements and complexity vary depending on factors such as technical maturity and the context in which AI is deployed.”
The new document is coming more than two years after DoD adopted its Ethical principles of AI and just over a year after it published its RAI memorandum, which guided the department’s approach to RAI. The new strategy comprises a set of “Principles”: RAI Governance, Warfighter Trust, AI Product and Procurement Lifecycle, Requirements Validation, RAI Ecosystem and AI Staff.
“It is imperative that the Department of Defense adopt responsible behaviors, processes and goals and implement them in a manner that reflects the Department’s commitment to its AI ethics,” the strategy reads. “Failure to use AI responsibly puts our war fighters, the public and our partnerships at risk.”
Each principle is accompanied by working lines, corresponding offices with primary responsibility and the estimated timeframe for implementation. The recently established Chief Data and AI Office will serve as the lead for the implementation of RAI.
TIED TOGETHER: Lyft Exec Craig Martell Tapped as Pentagon AI Chief: Exclusive Interview
In accordance with the principle of the RAI requirements, the CDAO, in coordination with the DoD component, directs – the Office of the Assistant to the Secretary of Defense for Privacy, Civil Liberties and Transparency; the common staff; and the military departments — will create a repository of AI-related common use cases, mission domains, and system architectures to “facilitate reusability.”
The CDAO will also develop an acquisition toolkit “that draws on best practices and innovative research from the DoD corporate, industry and academic community, as well as commercially available technology where appropriate.” The office will develop the toolkit in coordination with the Offices of the Undersecretary for Research and Technology and Procurement and Conservation.
The toolkit itself will provide a set of operationally relevant RAI-related assessment criteria, guidance on how industry can meet the DoD’s AI Code of Ethics, and “standard AI contract language that includes clauses for: independent government.” [test and evaluation] of AI capabilities, methods of immediate remediation when vendor-provided AI capabilities cannot be used in accordance with the DoD AI Code of Ethics, requesting training and documentation from vendors, performance monitoring of AI capabilities and appropriate data delivery, and -rights” and all other relevant resources.
The Assistant Secretary of Defense for Legislative Affairs at the CDAO will also develop a department-wide legislative strategy to “ensure appropriate collaboration with the CDAO and consistent messaging, technical assistance and advocacy to Congress.”
In another direction, the CDAO and the Office of the Undersecretary for Research and Technology will be responsible for providing the White House Office of the National AI Initiative with a prioritized list of research gaps in RAI-related areas in order to attract funding from the to promote National Institute of Standards and Technology, Department of Education and National Science Foundation.
As for the AI workforce, the strategy outlines efforts to strengthen it: Develop a mechanism to identify and track AI expertise across the DoD by leveraging existing coding efforts and developing standardized personal coding mechanisms; Conducting a gap analysis to determine if additional skills are required to successfully implement RAI; and other efforts to recruit and retain AI experts.
Aside from the workforce efforts outlined in the RAI Strategy and Implementation Pathway, the CDAO and Department of Defense Chief Information Officer John Sherman recently announced that their offices are developing a new digital workforce strategy that will be critical to reaching the talent that the… CDAO office will need.
Ultimately, the Pentagon’s target end state for RAI, according to its strategy, is trust. To achieve this desired end state, DoD cannot rely solely on technological advances.
“Key trustworthiness factors also include the ability to demonstrate a reliable governance structure and the provision of appropriate training and education to the workforce,” the strategy reads. “These efforts will help foster an appropriate level of trust so that the workforce no longer views AI as a mysterious and incomprehensible technology, but rather understands the capabilities and limitations of this widely used and accepted technology.”