Corin Stone, Washington College of Law
Corin Stone is a Scholar-in-Residence and Adjunct Professor at the Washington College of Law. Stone is on leave from the Office of the Director of National Intelligence (ODNI) where, until August 2020, she served as the Deputy Director of National Intelligence for Strategy & Engagement, leading Intelligence Community (IC) initiatives on artificial intelligence, among other key responsibilities. From 2014-2017, Ms. Stone served as the Executive Director of the National Security Agency (NSA).
(Editor’s Note: This article was first published by our friends at Just Security and is the third in a series that is diving into the foundational barriers to the broad integration of AI in the IC – culture, budget, acquisition, risk, and oversight.)
OPINION — As I have written earlier, there is widespread bipartisan support for radically improving the nation’s ability to take advantage of artificial intelligence (AI). For the Intelligence Community (IC), that means using AI to more quickly, easily, and accurately analyze increasing volumes of data to produce critical foreign intelligence that can warn of and help defuse national security threats, among other things. To do that, the IC will have to partner closely with the private sector, where significant AI development occurs. But despite the billions of dollars that may ultimately flow toward this goal, there are basic hurdles the IC still must overcome to successfully transition and integrate AI into the community at speed and scale.
Among the top hurdles are the U.S. government’s slow, inflexible, and complex budget and acquisition processes. The IC’s rigid budget process follows the standard three-year cycle for the government, which means it takes years to incorporate a new program and requires confident forecasting of the future. Once a program overcomes the necessary hurdles to be included in a budget, it must follow a complex sequence of regulations to issue and manage a contract for the actual goods or services needed. These budget and acquisition processes are often considered separately as they are distinct, but I treat them together because they are closely related and inextricably intertwined in terms of the government’s purchasing of technology.
Importantly, these processes were not intended to obstruct progress; they were designed to ensure cautious and responsible spending, and for good reason. Congress, with its power of the purse, and the Office of Management and Budget (OMB), as the executive branch’s chief budget authority, have the solemn duty to ensure wise and careful use of taxpayer dollars. And their roles in this regard are vital to the U.S. government’s ability to function.
Unfortunately, despite the best of intentions, as noted by some in Congress itself, the budget process has become so “cumbersome, frustrating, and ineffective” that it has weakened the power of the purse and Congress’ capacity to govern. And when complicated acquisition processes are layered on top of the budget process, the result is a spider web of confusion and difficulty for anyone trying to navigate them.
The Need for Speed … and Flexibility and Simplicity
As currently constructed, government budget and acquisition processes cause numerous inefficiencies for the purchase of AI capabilities, negatively impacting three critical areas in particular: speed, flexibility, and simplicity. When it comes to speed and flexibility, the following difficulties jump out:
- The executive branch has a methodical and deliberate three-year budget cycle that calls for defined and steady requirements at the beginning of the cycle. Changing the requirements at any point along the way is difficult and time-consuming.
- The IC’s budgeting processes require that IC spending fit into a series of discrete sequential steps, represented by budget categories like research, development, procurement, or sustainment. Funds are not quickly or easily spent across these categories.
- Most appropriations expire at the end of each fiscal year, which means programs must develop early on, and precisely execute, detailed spending plans or lose the unspent funds at the end of one year.
- Government agencies expend significant time creating detailed Statements of Work (SOWs) that describe contract requirements. Standard contract vehicles do not support evolving requirements, and companies are evaluated over the life of the contract based on strict compliance with the original SOW created years earlier.
These rules make sense in the abstract and result from well-intentioned attempts to buy down the risk of loss or failure and promote accountability and transparency. They require the customer to know with clarity and certainty the solution it seeks in advance of investment and they narrowly limit the customer’s ability to change the plan or hastily implement it. These rules are not unreasonably problematic for the purchase of items like satellites or airplanes, the requirements for which probably should not and will not significantly change over the course of many years.
However, because AI technology is still maturing and the capabilities themselves are always adapting, developing, and adding new functionality, the rules above have become major obstacles to the quick integration of AI across the IC. First, AI requirements defined with specificity years in advance of acquisition – whether in the budget or in a statement of work – are obsolete by the time the technology is delivered. Second, as AI evolves there is often not a clear delineation between research, development, procurement, and sustainment of the technology – it continuously flows back and forth across these categories in very compressed timelines. Third, it is difficult to predict the timing of AI breakthroughs, related new requirements, and funding impacts, so money might not be spent as quickly as expected and could be lost at the end of the fiscal year. Taken together, these processes are inefficient and disruptive, cause confusion and delay, and discourage engagement from small businesses, which have neither the time nor the resources to wait years to complete a contract or to navigate laborious, uncertain processes.
Engage personally with experts on Artificial Intelligence and national security at The Cipher Brief Threat Conference October 24-26. If you are an actively working in the national security field, we invite you to apply to attend. Seats are limited.
Simply put, modern practices for fielding AI have outpaced the IC’s decades-old approach to budgeting and acquisition. That AI solutions are constantly evolving, learning, and improving both undermines the IC’s ability to prescribe a specific solution and, in fact, incentivizes the IC to allow the solution to evolve with the technology. The lack of flexibility and speed in how the IC manages and spends money and acquires goods and services is a core problem when it comes to fully incorporating AI into the IC’s toolkit.
Even while we introduce more speed and agility into these processes, however, the government must continue to ensure careful, intentional, and appropriate spending of taxpayer dollars. The adoption of an IC risk framework and modest changes to congressional oversight engagements, which I address in upcoming articles, will help regulate these AI activities in the spirit of the original intent of the budget and acquisition rules.
As for the lack of simplicity, the individually complex budget and acquisition rules are together a labyrinth of requirements, regulations, and processes that even long-time professionals have trouble navigating. In addition:
- There is no quick or simple way for practitioners to keep current with frequent changes in acquisition rules.
- The IC has a distributed approach that allows each element to use its various acquisition authorities independently rather than cohesively, increasing confusion across agency lines.
- Despite the many federal acquisition courses aimed at demystifying the process, there is little connection among educational programs, no clear path for IC officers to participate, and no reward for doing so.
The complexity of the budget and acquisition rules compounds the problems with speed and flexibility, and as more flexibility is introduced to support AI integration, it is even more critical that acquisition professionals be knowledgeable and comfortable with the tools and levers they must use to appropriately manage and oversee contracts.
Impactful Solutions: A Target Rich Environment
Many of these problems are not new; indeed, they have been highlighted and studied often over the past few years in an effort to enable the Department of Defense (DOD) and the IC to more quickly and easily take advantage of emerging technology. But to date, DOD has made only modest gains and the IC is even further behind. While there are hundreds of reforms that could ease these difficulties, narrowing and prioritizing proposed solutions will have a more immediate impact. Moreover, significant change is more likely to be broadly embraced if the IC first proves its ability to successfully implement needed reforms on a smaller scale. The following actions by the executive and legislative branches – some tactical and some strategic – would be powerful steps to ease and speed the transition of AI capabilities into the IC.
Statements of Objectives
A small but important first step to deal with the slow and rigid acquisition process is to encourage the use of Statements of Objectives (SOO) instead of SOWs, when appropriate. As mentioned, SOWs set forth defined project activities, deliverables, requirements, and timelines, which are used to measure contractor progress and success. SOWs make sense when the government understands with precision exactly what is needed from the contractor and how it should be achieved.
SOOs, on the other hand, are more appropriate when the strategic outcome and objectives are clear, but the steps to achieve them are less so. They describe “what” without dictating “how,” thereby encouraging and empowering industry to propose innovative solutions. SOOs also create clarity about what is important to the government, leading companies to focus less on aggressively low pricing of specific requirements and more on meeting the ultimate outcomes in creative ways that align with a company’s strengths. This approach requires knowledgeable acquisition officers as part of the government team, as described below, to ensure the contract includes reasonable milestones and decision points to keep the budget within acceptable levels.
New Authorities for the IC
Two new authorities would help the IC speed and scale its use of AI capabilities: Other Transaction Authority (OTA) and Commercial Solutions Openings (CSO). Other Transaction Authority allows specific types of transactions to be completed outside of the traditional federal laws and regulations that apply to standard government procurement contracts, providing significantly more speed, flexibility, and accessibility than traditional contracts. While OTA is limited in scope and not a silver bullet for all acquisition problems, OTA has been used to good effect since 1990 by the Defense Advanced Research Projects Activity (DARPA), DOD’s over-the-horizon research and development organization, among others.
CSOs are a simplified and relatively quick solicitation method to award firm fixed price contracts up to $100 million. CSOs can be used to acquire innovative commercial items, technologies, or services that close capability gaps or provide technological advances through an open call for proposals that provide offerors the opportunity to respond with technical solutions of their own choosing to a broadly defined area of government interest. CSOs are considered competitively awarded regardless of how many offerors respond.
Both OTA and CSO authority should be immediately granted to the IC to improve the speed and flexibility with which the IC can acquire and transition AI into the IC.
The predictive nature of the IC’s work and the need to forecast outcomes means the IC must be able to acquire AI at the point of need, aligned to the threat. Waiting several years to acquire AI undermines the IC’s ability to fulfill its purpose. But with speed comes added risk that new capabilities might fail. Therefore, the IC should create an isolated unclassified sandbox, not connected to operational systems, in which potential IC customers could test and evaluate new capabilities alongside developers in weeks-to-months, rather than years. Congress should provide the IC with the ability to purchase software quickly for test and evaluation purposes only to buy down the risk that a rapid acquisition would result in total failure. The sandbox process would allow the IC to test products, consider adjustments, and engage with developers early on, increasing the likelihood of success.
Single Appropriation for Software
DOD has a pilot program that funds software as a single budget item – allowing the same money to be used for research, production, operations, and sustainment – to improve and speed software’s unique development cycle. AI, being largely software, is an important beneficiary of this pilot. Despite much of the IC also being part of DOD, IC-specific activities do not fall within this pilot. Extending DOD’s pilot to the IC would not only speed the IC’s acquisition of AI, but it would also increase interoperability and compatibility of IC and DOD projects.
Congress should reconsider the annual expiration of funds as a control lever for AI. Congress already routinely provides no-year funding when it makes sense to do so. In the case of AI, no-year funds would allow the evolution of capabilities without arbitrary deadlines, drive more thoughtful spending throughout the lifecycle of the project, and eliminate the additional overhead required to manage the expiration of funds annually. Recognizing the longer-term nature of this proposal, however, the executive branch also must seek shorter-term solutions in the interim.
A less-preferable alternative is to seek two-year funding for AI. Congress has a long history of proposing biennial budgeting for all government activities. Even without a biennial budget, Congress has already provided nearly a quarter of the federal budget with two-year funding. While two-year funding is not a perfect answer in the context of AI, it would at a minimum discourage parties from rushing to outcomes or artificially burning through money at the end of the first fiscal year and would provide additional time to fulfill the contract. This is presumably why DOD recently created a new budget activity under their Research, Development, Test and Evaluation (RDT&E) category, which is typically available for two years, for “software and digital technology pilot programs.”
AI Technology Fund
Congress should establish an IC AI Technology Fund (AITF) to provide kick-starter funds for priority community AI efforts and enable more flexibility to get those projects off the ground. To be successful, the AITF must have no-year funds, appropriated as a single appropriation, without limits on usage throughout the acquisition lifecycle. The AITF’s flexibility and simplicity would incentivize increased engagement by small businesses, better allowing the IC to tap into the diversity of the marketplace, and would support and speed the delivery of priority AI capabilities to IC mission users.
To quickly take advantage of private sector AI efforts at scale, the IC must better understand the market and more easily engage directly with the private sector. To do so, the IC should create an ICWERX, modeled after AFWERX, an Air Force innovation organization that drives agile public-private sector collaboration to quickly leverage and develop cutting-edge technology for the Air Force. AFWERX aggressively uses innovative, flexible, and speedy procurement mechanisms like OTA and the Small Business Innovation Research and Small Business Technology Transfer programs (SBIR/STTR) to improve the acquisition process and encourage engagement from small businesses. AFWERX is staffed by acquisition and market research experts who are comfortable using those authorities and understand the market. While the IC’s needs are not identical, an ICWERX could serve as an accessible “front door” for prospective partners and vendors, and enable the IC to more quickly leverage and scale cutting-edge AI.
De-mystify Current Authorities
While there is much complaining about a lack of flexible authorities in the IC (and a real need for legal reform), there is flexibility in existing rules that has not been fully utilized. The IC has not prioritized the development or hiring of people with the necessary government acquisition and contracts expertise, so there are insufficient officers who know how to use the existing authorities and those who do are overworked and undervalued. The IC must redouble its efforts to increase its expertise in, and support the use of, these flexibilities in several ways.
First, the IC should create formal partnerships and increase engagement with existing U.S. government experts. The General Services Administration’s Technology Transformation Services (TTS) and FEDSIM, for example, work across the federal government to build innovative acquisition solutions and help agencies more quickly adopt AI. In addition, DOD’s Joint AI Center has built significant acquisition expertise that the IC must better leverage. The IC also should increase joint duty rotations in this area to better integrate and impart acquisition expertise across the IC.
Second, the IC must prioritize training and education of acquisition professionals. And while deep acquisition expertise is not necessary for everyone, it is important for lawyers, operators, technologists, and innovators to have a reasonable understanding of the acquisition rules, and the role they each play in getting to successful outcomes throughout the process. Collaboration and understanding across these professions and up and down the chain of command will result in more cohesive, speedy, and effective outcomes.
To that end, the Office of the Director of National Intelligence (ODNI) should work with the many existing government acquisition education programs, as well as the National Intelligence University, to develop paths for IC officers to grow their understanding of and ability to navigate and successfully use acquisition rules. The ODNI also should strengthen continuing education requirements and create incentive pay for acquisition professionals.
Third, the IC should prioritize and use direct hire authority to recruit experts in government acquisition, to include a mix of senior term-limited hires and junior permanent employees with room to grow and the opportunity for a long career in the IC. Such a strategy would allow the IC to quickly tackle the current AI acquisition challenges and build a bench of in-house expertise.
Finally, practitioners should have an easily accessible reference book to more quickly discover relevant authorities, understand how to use them, and find community experts. A few years ago, the ODNI led the creation of an IC Acquisition Playbook, which describes common IC acquisition authorities, practices, and usages. The ODNI should further develop and disseminate this Playbook as a quick win for the IC.
To encourage creative and innovative acquisition practices, as well as interdisciplinary collaboration, the IC must align incentives with desired outcomes and create in acquisition professionals a vested interest in the success of the contract. Acquisition officers today are often brought into projects only in transactional ways, when contracts must be completed or money must be obligated, for example. They are rarely engaged early as part of a project team, so they are not part of developing the solutions and have minimal investment in the project’s success. Reinforcing this, acquisition professionals are evaluated primarily on the amount of money they obligate by the end of the fiscal year, rather than on the success of a project.
Therefore, to start, project teams should be required to engage acquisition officers early and often, both to seek their advice and to ensure they have a good understanding of the project’s goals. In addition, evaluation standards for acquisition officers should incorporate effective engagement and collaboration with stakeholders, consideration of creative alternatives and options, and delivery of mission outcomes. If an officer uses innovative practices that fail, that officer also should be evaluated on what they learned from the experience that may inform future success.
Lastly, the ODNI should reinvigorate and highlight the IC acquisition awards to publicly reward desired behavior, and acquisition professionals should be included in IC mission team awards as a recognition of their impact on the ultimate success of the mission.
Between the government’s rigid budget and acquisition processes and confusion about how to apply them, there is very little ability for the IC to take advantage of a fast-moving field that produces new and updated technology daily. Tackling these issues through the handful of priority actions set forth above will begin to drive the critical shift away from the IC’s traditional, linear processes to the more dynamic approaches the IC needs to speed and transform the way it purchases, integrates, and manages the use of AI.
Read more expert-driven national security insights, perspective and analysis in The Cipher Brief