LAUREN KAHN and MICHAEL HOROWITZ,
The Department of Defense is considering organizational changes designed to create a more integrated approach to data and artificial intelligence, including the creation of a Chief Data and Artificial Intelligence Officer. If the reorganization occurs, the CDAO will oversee several pre-existing offices, including the office of the Chief Data Officer, the Joint Artificial Intelligence Center, and the Defense Digital Service. Consolidated oversight through creating an empowered CDAO could help ensure DoD has the tools it needs to excel and ensure U.S. defense innovation leadership moving forward.
Technology leadership requires data and AI leadership, and right now DoD’s data and AI efforts are splintered. For example, according to Govini, a decision science company based in Virginia, at least 15 separate institutions within DoD invest to some extent in artificial intelligence, AI adjacent technologies, foundational enabling capabilities for AI, or programs that use AI during development. Each with its own separate processes, data, code, and programs. There is not enough structured coordination between them or oversight through a central “AI hub.” This undermines the ability of the United States to lead in emerging technologies and defend DoD’s networks.
Would a CDAO have the ability to coordinate efforts across DoD effectively and the authority to compile and share data in cases where military services or other components could be hesitant? Moreover, it will be critical for ensuring the organization can maintain the focus necessary for the JAIC and DDS to excel while also institutionalizing their efforts in a way that positions them to succeed in an enormous defense bureaucracy. It will be paramount for real institution-building to occur.
Why Bring Data and AI Together?
Adopting emerging technologies requires experimenting with the institutional structures necessary to take advantage of them. The consolidation of the office of the CDO, the JAIC, and the DDS offices under the CDAO represents a potential moment of organizational growth for DoD when it comes to taking data and emerging technology seriously. The Chief Data Officer is responsible for data management and coordination across the Department. The JAIC was created to help DoD enable and implement uses of artificial intelligence. The Defense Digital Service is a data science strike team designed to handle data and security issues. These formerly disconnected entities all control an important piece of the data and emerging technology puzzle—bringing them together will help to propel the DoD into the future.
Today, data within the DoD is too siloed, noisy, and unlabeled, making usage difficult and expensive. Data is “food for AI,” and data preparation and management are the most time-consuming parts of developing and training algorithms. Therefore, having quality and easy-to-use data will be critical for the DoD’s ability to successfully implement AI-enabled systems and processes, especially given most algorithms are open-source. An integrated reorganization could help to elevate the importance of data within DoD, assure a level of internal consistency, and facilitate resource sharing which will help set it up for success in adopting AI and other emerging technologies—heavily reliant on data—down the line.
Organizational integration under the CDAO will also offer potential advantages for each of the constituent parts. The creation of the JAIC followed best practices from military innovation and business innovation literature surrounding the need to create spinoff or separate sub-organizations to value the potential of emerging technologies. The JAIC has become a large organization working on a number of projects at any one time, raising questions of prioritization. Earlier in 2021, the JAIC became a direct report to the Deputy Secretary of Defense. Previously, it reported to the Chief Information Officer. Some might interpret the proposed reorganization as evidence DoD is taking AI less seriously since the JAIC would now report to the CDAO, who in turn reports to the Deputy Secretary. But the JAIC is torn between being a developer of algorithms itself and being an enabler that helps the military services figure out how to develop and implement algorithms within relevant military programs. The JAIC also has lacked the authority to compel the military services and other institutions to collaborate. A tight linkage between DoD’s data and AI hubs will give DoD’s algorithmic efforts better access to necessary data to succeed.
A CDAO could also better position DoD to fully implement the May 2021 Data Directive from the Deputy Secretary of Defense, which aims to maximize data availability and create interfaces that help DoD access the huge amount of data it controls. This will help DoD better understand itself, which in turn will make DoD more effective. The consolidation and standardization of different streams of data, and its flow through one central node that can then redistribute it, will lead to data that is more usable and will facilitate better data sharing practices across DoD.
Being placed under the umbrella of the CDAO will also help the DDS increase its reach throughout DoD. DDS is an inherently disruptive organization—it focuses on bringing in top data and technology talent to tackle hard problems such as drone detection and bug bounty programs to close cyber vulnerabilities. Given the talent pipeline for DDS comes more from the private sector via short-term-limited appointments, DDS works best when tasked with specific projects. Closer integration with DoD’s data hub will help ensure DDS has the tools necessary to succeed, as well as generate synergy between DDS and broader DoD AI efforts.
OVERCOMING IMPLEMENTATION CHALLENGES
The proposed reorganization, no matter the intentions of senior leaders, could face bureaucratic or other headwinds that undermine the goal of consolidated AI, data, and digital leadership. The organization needs the authority to get DoD components, especially the military services, to share data from databases never designed for sharing. Even with the DoD Data Directive in hand, the CDAO will need the strong support of the Deputy Secretary to succeed. The reorganization appears necessary, in part, because existing institutions, despite good intentions, do not have the focus to coordinate DoD efforts in data and AI. That lack of focus, if not addressed, could undermine the authority of the CDAO and prevent it from bringing together DoD’s efforts. DoD’s leadership must not lose sight of the critical importance of AI to DoD’s future—for its internal activities and future military operations. The melding of DoD’s data and AI missions should strengthen Department AI efforts, rather than lead to their minimization. Finally, a CDAO would need leadership and staffing that combines knowledge of substantive policy issues surrounding military applications of data and AI with technical knowledge. Insufficient technical knowledge could risk the CDAO selling smoke, while insufficient policy knowledge could mean the CDAO does not advance efforts in the most important areas for U.S. defense capabilities.
The potential reorganization of DoD department-wide AI and data efforts provides an opportunity for more integrated and streamlined emerging technology leadership that more effectively connects to DoD capabilities. Close relationships with other technology accelerators, whether in Research & Engineering or in the military services, will be essential. Like all reorganizations, this one will ultimately succeed or fail based in part on the support it receives from senior DoD leadership, and the ability to get buy-in from the constituent organizations and surrounding bureaucracy.
No comments:
Post a Comment