The Defense Information Systems Agency (DISA) announced a new prototype effort that will use machine learning and artificial intelligence to improve decision-making and timeliness when it comes to the security clearance process.
The federal government faces the problem of processing security clearances accurately and quickly. “When I listen to the sponsors of this program, the bottom line is getting people clear to work in their jobs faster,” Mr. Terry Carpenter, Program Executive Officer at the National Background Investigation Service (NBIS), said at a reporter roundtable. “We’re not doing that well today.”
Processing security clearances timely and accurately is critical for expediting the onboarding process, retaining more government workers and preventing security risks. A $75 million other transaction agreement (OTA) is a step forward to an innovative NBIS IT system, which is a case-control and management system, and a better security clearance personnel vetting process.
DISA partnered with the Defense Security Service (DSS), which is the Defense Department’s security oversight provider, to award the OTA to Perspecta Enterprise Solutions of Herndon, Virginia on May 14. There are five non-traditional subcontractors also part of the award; the federal government has never done business with three of these subcontractors. An OTA is a contract that is generally exempt from federal procurement laws and regulations, according to the Congressional Research Service. For this reason, OTAs can prove controversial. However, government officials are given greater flexibility to shape agreements, and they can amend clauses or requirements that are necessary for traditional procurements.
DSS and DISA awarded the OTA to further enhance the foundations of NBIS to form a prototype that would allow for data reuse, analysis, and security across multiple missions. The clearance model has gone unchanged for many years, but now there’s a newfound commitment to evolve and improve the model based on new technology and new risk categories.
This award follows the $49 million awarded to Enterprise Solutions last June to build the NBIS Investigation Management (IM) Shared Service and integrate multiple disparate systems into one enterprise. The two awards concern different projects- the award decided this month looks to build security clearance architecture.
From an acquisitions standpoint, something of this magnitude would usually take one to two years to complete; this took agency leaders five to six months, from phase one to phase three, which was the awarding of the OTA. Already, the agencies involved are establishing a precedent of working thoroughly in a fraction of the usual time.
The prototype especially considers the two main areas of focus: data capabilities bolstered by AI, like machine learning, and capabilities to develop, test and secure security clearance vetting with DevSecOps automation.
This new foundation will allow DSS to transform vetting processes and organize data quickly and more efficiently, with special regard for factors such as current laws, rules and guidelines that can influence decision-making. Thus, DSS can make better-informed decisions in less time, and create unique innovations around security controls.
While leaders at DSS and DISA are hesitant to put a time frame on an expedited clearance process, they are aiming for a full transformation of their business process from end-to-end.
The government is the lead systems integrator on the OTA. Agency leaders wanted to maintain the robustness expected of federal services and make sure that new security clearance processes work with cloud architecture.
The general criteria required for security clearances include criminal, financial, credit and public records. In new transformation policies under discussion, there has been talk of using publicly available information, like social media, but those policies have not been finalized.
The security clearance process fundamentally consists of three different parts. The first is the initiation process, which involves engagement with the subject and collecting information. Then comes the investigation, followed by the adjudication.
Agencies are looking to continuous vetting to minimize the labor force that has to go out and knock on doors to get information. Currently, the validation process in place involves reinvestigations every five years. However, if a set of data allows for more confidence and accuracy, it also leads to greater trust in subsequent steps of analysis. The shift to continuous as opposed to periodic validation follows attempts to elevate confidence in the data and the clearances that are issued.
The bottom line is that agencies need to get people to work in their jobs faster. Whether those people are found in federal agencies, or in private partner companies, the process for getting people to work should be easier, especially as new or potential hires attempt to seek out meaningful work.
Ultimately, this contract lays the groundwork for helping expedite the security clearance process and enables people to get to work faster.
AI and machine learning could have such a huge impact on this process! I hope it works. I often hear that the security clearance road is pretty rocky for lots of potential govies.