Last fall, leaders from the Joint Artificial Intelligence Center began developing the JAIC Data Governance Council with the goal of establishing an enterprise-wide data governance framework. This framework is specifically designed to help JAIC users manage various forms of data at every stage of project development.
Data governance is the process of managing the availability, usability, integrity, and security of data being used in large enterprise systems. This is a critical need for an organization such as the JAIC that must ingest and organize large volumes of data on a regular basis.
“The data governance policies we put in place will increase effectiveness and efficiency, reduce complexity, and enable better data sharing.”
That’s where the DGC comes into play.
The council, which is still in its initial phase of development, is on target to be fully functional by September 2020. The DGC is made up of JAIC representatives from the Joint Common Foundation, along with the JAIC’s product, policy, and mission teams. The council’s overall mission is to make sure that the JAIC properly securing, organizing, storing, and managing its growing collection of data assets on the JCF. This work will impact not only the JAIC but also the broader DoD enterprise.
“We’re dealing with such scalability, so we need to make sure that we’re minimizing complexity,” said Sylvia Mapes, data lead for the JCF and a member of the DGC. “The ways of doing business in a medium-sized entity are not going to work in the DoD, so the data governance policies we put in place will increase effectiveness and efficiency, reduce complexity, and enable better data sharing.”
In the interim, the JAIC has established a working group to develop and finalize the council’s long-term strategy and structure. The group is also focused on sharing its work across the federal AI community to leverage best practices and garner community buy-in.
“The Data Governance working group is tackling a number of workstreams, such as translating information sharing agreements into technical requirements or capturing important data so it can be published to the AI product development ecosystem,” explains Bryan Lane, director of data and AI for the General Services Administration Centers of Excellence and a lead member of the DGC. “These are not trivial tasks, and as we work together as a group to complete them, those lessons can be shared across the DoD, enabling AI adoption.”
The involvement of the JCF will be especially important, as it will be responsible for determining how data will be ingested and secured inside the JAIC. In addition, the JCF representative in the council will also “serve as the lead technical advisor on all JAIC data lifecycle management activities and establish data management processes that enable the technical implementation of data discovery, access, and use.”
The JCF lead will be responsible for data ingest, remote data access, data dissemination, and interoperability across federated systems and will work closely with the existing Mission Initiatives. Even with this significant oversight, the JCF will still rely on “data stewards,” designated points-of-contact for the council within each MI, to provide specific direction on how, when, and where data is used to advance their AI project.
The work of the JAIC DGC will address the following objectives:
- Easing the sharing of data within the JAIC.
- Creating better AI products by increasing efficiency.
- Ensuring data consistency, reliability, and repeatability.
- Reducing complexity.
- Bringing key stakeholders to the table to make decisions about data.
- Maximizing money and manpower for AI/ML project development.
Join The JAIC on Twitter