Artificial Intelligence in the U.S. Government
A Mission-Oriented Journey from Dark Data to Open Data and Data Innovation
Sponsored by Sov.ai
Abstract
The United States Federal Data Strategy is central to the vision for modernizing the Federal Government and driving data innovation towards enabling federal agencies to deliver mission. AI provides an extraordinary opportunity to engage in the process of shaping enterprise-level data strategies to achieve mission outcomes.
The Data Strategy
The United States (U.S.) Federal Data Strategy is central to the vision for modernizing the Federal Government and driving data innovation towards enabling federal agencies to deliver mission outcomes, increase effectiveness, and provide excellent service on behalf of the American people. The 2020 Action Plan specifies initial actions to be taken by agencies to establish a solid foundation that will support implementation of the strategy over the next decade. This is a central component in promoting the adoption of Artificial Intelligence (AI) within the federal public sector.
While several agencies have ventured in the AI space, they have their own specific analytic goals and requirements. Their current and planned capabilities are at different stages in the AI spectrum. However, while use cases may differ, the underlying requirements are fundamentally similar and include the maturity of the data infrastructure, the capability to support the data supply chain, and the leadership engagement to drive AI game changing methodologies and solutions. AI capabilities are to be leveraged at the technical, organizational, and operational levels.
The U.S. Commerce Department’s National Technical Information Service (NTIS) delivers a Fed-to-Fed framework for data science innovation through precedent-setting partnerships with industry, universities, and non-profits. It provides a unique pathway for federal agencies towards innovation and digital transformation to achieve mission outcomes effectively and efficiently.
The Data Challenge
The COVID-19 pandemics has exacerbated the need for government to be able to anticipate and prevent problems rather than just react to them. One of the most critical components of any operational entity is its data. To leverage the necessary data to support operations and mission-critical applications, data discovery, collection, aggregation, access, analysis, dissemination, and security are crucial. This requires timely access to integrated data from disparate sources and the capability for cross-organizational intelligence sharing internally and among government agencies, while maintaining the highest standards of data governance, data quality, and data security.
Over the years, the government has collected and stored huge data sets in support of specific applications typically executed in legacy systems. To unleash the power of data analytics and artificial intelligence, a multipronged data strategy is needed. It should be geared towards eliminating data silos, migrating from and eventually retiring legacy systems, standardizing data collection and ingestion, and integrating data sources both within and across agencies.
The U.S. Federal Strategy aims to develop a government-wide data strategy along with an action plan for implementation. The objective of the Federal Data Strategy is to fully leverage the value of federal data for mission, service, and the public good by guiding the Federal Government in practicing ethical governance, conscious design, and a learning culture. The evidence suggests that we are amid an unprecedented digital transformation fueled by the intersection of several emerging technologies, including cloud computing, big data, Internet of Things (IoT), Blockchain, and AI.
This provides an extraordinary opportunity for public or private enterprises to engage in the process of shaping enterprise-level data strategies to develop data-driven and innovative methodologies to support business processes, empower evidence-based decision making, and produce and operationalize modern and data-driven infrastructures. For federal agencies, migrating to an integrated, analytic data platform to fully leverage data as a strategic asset to achieve mission outcomes is no small task. However, like any technical advancement, modernizing and future-proofing the enterprise critical infrastructure is an essential investment that can uncover valuable untapped relationships within the data. The potential benefits include evidence-based decision making, improved business processes, enhanced data governance and security, compliance with regulations on data retention, sensitivity, and privacy.
The Data Science Hierarchy of Needs is a pyramid showing what is needed to support data-driven applications. A quintessential version is shown in Figure 1 (The AI hierarchy of needs, 2017). The base of the pyramid involves capturing all the relevant data that need to be aggregated, curated, and harmonized to eventually support intelligence in the production system. Most government organizations have gaps in the base of the pyramid. Ultimately, the goal of the U.S. Federal Data Strategy is to prepare government agencies to derive value from data while supporting robust data governance and protecting security, privacy, and confidentiality. It is vital to empower government agencies to analyze, augment, aggregate, correlate, and consolidate growing volumes of data in new ways that lead to deeper insight and greater efficiency. This will fuel the drive for innovation and bolster productivity across multiple sectors to consolidate the U.S. economic competitiveness.
AI is fundamentally rooted in data. Investment in data capabilities is essentially an investment in AI. Any application of AI and Machine Learning (ML) will only be as good as the quality of data collected. The transformative potential of AI includes the practice of aggregating and analyzing data from many sources to derive insights and to anticipate future outcomes. It would allow agencies to target likely problems before they erupt into crises. It can be applied in a wide range of areas including defense, security, health care, and human services, among others. Ultimately, every agency needs to aim for excellence in data management and analytics, in which data is logically stored to allow full connectiveness and support for the whole range of analytics, from descriptive and diagnostic, to predictive and prescriptive.
The Innovation Challenge
Driving the transformation of the federal enterprise through innovation is essential to achieve mission outcome. Emerging and innovative technologies provide unprecedented opportunities to address mission-critical challenges and provide customer-driven services delivered efficiently, effectively, faster, and at lower cost. Succeeding in this effort requires a robust, healthy, transparent ecosystem of robust data and AI capabilities to support the entire business process from the genesis of the data to the experience of the end user. This includes a unified data management platform and tools and techniques for collecting, analyzing, disseminating, and managing large volumes of data at high velocity with different data types and flows.
However, innovation is a hard journey. It requires from any organization to transform itself and change its processes, services, business models while integrating risk and change management efforts. To succeed, a federal agency should proceed with an operating framework that enables it to explore innovative projects with tangible, immediate impact congruent with its national mission. This journey includes essential ingredients such as leveraging a modern infrastructure to sustain emerging technologies and a data platform for consistent and interoperable data collection, ingestion, and dissemination. U.S. government agencies should also prioritize involvement in AI standards efforts that are inclusive and accessible, open and transparent, consensus-based, globally relevant, and non-discriminatory (nist.gov, 2020).
The U.S. Federal Register indicates there are over 430 departments, agencies, and sub-agencies in the federal government. While some agencies are developing AI applications, the most successful ones are integrating AI into their operations to automate processes, improve quality, and increase efficiency. Rapid developments in AI have the potential to reduce the cost of core governance functions, improve the quality of decisions, and unleash the power of administrative data, thereby making government performance more efficient and effective. The assessment of the state of AI adoption within the federal government provides a widely different landscape. A snapshot of a survey conducted by the Government Business Council (GBC) in February and March 2019 on federal government organization’s AI priorities, challenges, and perspectives is shown in Figure 2 (Govexec.com, 2019).
Evidence-Based Decision Making
The U.S. Foundations for Evidence-Based Policymaking Act of 2018, which includes the Open, Public, Electronic and Necessary (OPEN) Government Data Act, sets the stage for major government-wide reforms for making data accessible and useful for decision-making. The Act defines how the government collects, publishes, and uses public data with the end goal of making evidence-based data more available for use in policymaking. The Act included key recommendations from the Commission on Evidence-Based Policymaking that will strengthen privacy protections, improve secure access to data, and enhance the government’s evidence-building capacity. The legislation provides the capacity and governance for evidence building activities and the use of data. The OPEN Government Data Act provides a sweeping, government-wide mandate for federal agencies to publish all their information as open data — using standardized, non-proprietary formats. It also incorporates recommendations from the U.S. Commission on Evidence-Based Policymaking to strengthen privacy and confidentiality of agency data and to adopt widely accepted strategies for securely analyzing existing data.
AI and associated algorithms support this new vision of data-driven government dealing with the vast amount of data that governmental agencies have access to. Artificial Intelligence techniques are increasingly extending and enriching decision support. Decision making is complicated and sometimes can be a critical task. Bias is a natural byproduct of the way human brains work. Data can carry over human bias and other types of bias. However, data-driven-AI and machine learning can be powerful decision-making instruments. They can augment the decision-making process by providing objective, valuable, usable, and shareable insights that far exceed current capabilities. AI techniques have been extensively used to support and enhance the quality of decision making and complex problem solving in different industries for many years through the exploitation of various types of machine intelligence, including natural language comprehension, robotics, expert systems, neural networks and machine learning.
It is essential for government agencies to develop capabilities to efficiently create rigorous evidence and routinely and securely use good information in government operations on which to base decisions and to construct effective evidence-based decision making. The Act requires for information to be accessible, understandable, useful, and relevant so it can inform agency activities and actions such as budgeting, program improvement, accountability, management, regulatory action and policy development. The American public wants a government that operates effectively and efficiently. However, this cannot be achieved without trust. To that end, the Act requires agencies to have evaluation officers to ensure rigor and transparency in the methodologies and statistical models that support decision making processes. On the other hand, early this year, the U.S. government published draft rules for the regulation of AI for public review and comment.
NTIS Framework
The Department of Commerce’s National Technical Information Service (NTIS) mission is to assist U.S. government agencies in leveraging data to achieve mission outcome and to become more effective and efficient to deliver citizen services. It is available to all Federal Agencies seeking an agile capacity to scale to meet critical mission data priorities. As a trusted Fed-to-Fed advisor, NTIS applies its unique joint venture partnership (JVP) authority (15 U.S. Code 3704b) leveraging leading, innovative companies to help federal agencies solve their complex data challenges.
Through the joint venture partner ecosystem, NTIS can bring in the nation’s top private sector companies, academic institutions, and non-profits to address national data centric challenges that require quick access to private sector ingenuity and expertise to achieve a capacity to scale and time to market. This Authority operates outside the U.S. Federal Acquisitions Regulation (FAR), which is the primary regulation for use by all executive agencies in their acquisition of supplies and services with appropriated funds. It enables the execution of Interagency Agreements to drive innovation in a truly flexible and Agile framework. The overall objective is to accelerate results across the Federal Government by leveraging data as a strategic asset to achieve mission outcomes.
The NTIS framework ethos is to drive innovation to impact mission. Innovation, by definition, is the introduction of something new. The framework promotes innovation by facilitating a diverse conversation between the private sector and the Federal Government at a much earlier stage of the problem definition process. This timely conversation is critical for driving innovation, and is difficult to achieve in a timely manner under the FAR. The first step is an ideation phase with the federal agency sponsoring the project. The ideation phase is all about learning about the agency mission, history, vision, strategy, and pain points from all stakeholder’s point of view. This initial discovery process starts with no preconceived notions or solutions in research of a problem. It is all about identifying and describing the right problem. As a result of this phase, a problem statement is developed in collaboration with the federal agency that defines the problem at hand, the project objective and critical success factors, the business requirements, and risk assessment factors.
In the conceptualization phase, the problem statement evolves in a project opportunity for the Joint Venture Partner(s) to propose their approach to deliver the vision and the plan that will actualize the agency’s vision through a series of transformative solutions, beginning with a concept of operations (CONOPS). The CONOPS is described as a concept of how the client operates as a modernized agency that has an innovative process, technology, and governance to manage the data. It will provide a framework for understanding the proposed functionality of the capability to be delivered, incorporating the data management lifecycle and the required governance, security, privacy, and other considerations. The vision described in the CONOPS will be actualized through a proof-of-concept. A proof-of-concept is defined as a visual representation of a series of capabilities that do not require software development — a proof-of-concept is simply a demonstration of capabilities to assist the client in envisioning future functionalities. The remaining effort must be focused on creating the prototype and use case. A prototype is defined as a capability that can be scaled and operationalized, e.g., on the client’s network or in the cloud, after all security requirements are addressed.
The JVP proposals go through a rigorous, transparent, fair, and equitable review process. Once the joint venture partner(s) selection and the interagency agreement are completed, the project kickoff immediately begins.
In summary, the NTIS framework is an excellent vehicle to develop a good problem statement that is human-centered and user-focused. By involving the agency early in the ideation phase, it allows to learn who your users are and, most importantly, their wants, needs, and pain-points. The empathy is transformed into an actionable problem statement. The other “secret sauce” of the model is to engage the private sector in the problem formulation phase, which turns out to be critical in ultimately solving the right problem. Finally, since NTIS operates outside the FAR, it enables a truly Agile development process inherently flexible to data-driven innovation, open to changing requirements over time, and adaptable to constant feedback from the end users. Most projects start within a 90-day process through an InterAgency Agreement with NTIS and the joint venture partners.
NTIS Use Cases
The core of NTIS mission is to assist federal agencies to leverage data as a strategic asset and harness AI and emerging technologies to achieve mission outcomes. Through its direct access to its joint venture partners’ expertise and know-how, the NTIS framework has been leveraged by multiple U.S. federal government agencies in several projects for a variety of mission outcomes across the federal sector, including the civilian, national security, defense, labor, commercial, economic, social, state, and healthcare sectors.
The NTIS framework has been used to tackle agency data challenges. One of the most common issue is the ability for cross intelligence and collaboration within and among agencies. Federal agencies with multiple bureaus, programs, and components present the challenges of inadequate data sharing, limited and distributed resources, and inefficient analytical tools to drive insights from data across the agency. The data challenge is exacerbated due to the fact that, although government has access to a lot of data, it was collected without AI in mind.
It is essential to establish robust data governance that ensures enriched data collection with specific parameters and in compliance with data standards and protection regulation. This needs to be conjointly carried out with a push for IT modernization and access to state-of-the-art computer hardware, frameworks, and software capabilities. The objective is to develop a real data enterprise where real-time information is used to inform decisions and operate from a single source of truth and an enhanced data maturity model. In this context, NTIS has helped design and build data analytics platforms to provide the capability to share information and foster knowledge management to meet mission demands.
NTIS has also engaged with several agencies to accelerate the process to access and analyze data to identify and target potential fraud schemes and areas of program waste and abuse. Through the leveraging of the NTIS framework, agencies can have access to an enterprise resources dashboard and Agile-delivered applications in a secure cloud-based platform. Capabilities include expanded business intelligence, allocation and coordination of the agency resources to improve decision-making, predictive capabilities agency wide to identify emerging issues through data and automated risk assessments, prioritization of oversight and enforcement resources, and optimization of resources and time to support key mission objectives.
A comprehensive and effective cybersecurity strategy is critical to secure federal systems, manage cybersecurity risks, and protect cyber critical infrastructure, privacy, and sensitive data. NTIS has provided support for several initiatives in this regard, including the enhancement of risk assessment methodologies to systematically measure, analyze and report on the risk exposure on different levels, including strategic, operational, transactional, and compliance.
AI is leveraged to enable agencies to be proactive, rather than reactive to prevent attacks before they happen and eliminate or minimize their effects. NTIS solutions include AI-enabled intelligent business automation, augmentation, and analytics capabilities that automate manual, high volume, error-prone, and repetitive tasks to support automated real-time risk management technology used to detect potential risks, analyze them, prioritize them based on the threat-level, and initiate alerts and actionable insights. Responses can be triggered automatically or escalated to human security experts.
Summary and Outlook
The confluence of transformative and game changing innovative and emerging technologies such as data science and AI provide an extraordinary opportunity for federal agencies to achieve mission outcomes, optimize business operations, and achieve quality customer service. The potential benefits affect all aspects of government work and mission. They include the creation of value at scale, reduction of time to market, optimization of procedures and processes, enhancement of data-driven decision making, improvement of the quality of services, and catalyzing economic growth and prosperity. However, there are several risk factors such as effectively managing change and disruption, readying the workforce, dealing with AI ethics and lack of explainability and transparency, and ensuring data privacy and quality.
Digital transformation is disruptive. It requires a full transformation of the essential body parts of the enterprise. The transformation starts at the core. In U.S. federal government departments and agencies, the mission is the central core. The path for AI adoption in sometimes risk-averse environment is to be able to deliver value that affects the mission. The many challenges that impede the adoption of AI and emerging technologies include a robust strategy and the buy-in to fund the modernization of systems, the development of data platforms, and the workforce to implement data-driven solutions. Is there a way to lay out a pathway towards data-driven innovation, evidence-based decision making, and mission effectiveness the federal ecosystem?
The NTIS framework provides a unique innovation value proposition for federal agencies:
- It is anchored in NTIS and the federal agency embarking on an innovation journey that leverages the agile methodology to continuously tailor the solution to customer needs and deliver innovation through proof points and transformative and scalable business operations.
- The private sector is engaged very early in the process (“pre-solicitation”) of problem formulation (not the solution) with the goal of understanding the mission and vision of the agency. A trusted relationship is established to discuss pain points and root causes, and collaboratively develop an initial problem statement.
- The NTIS approach is not an acquisition or procurement. The NTIS private sector venture authority offers federal agencies a unique business model to solve national/global data centric challenges. NTIS provides quick access to private sector ingenuity to achieve a capacity to scale and time to market/value, which is critical in ultimately solving the right problem while mitigating risk.
The U.S. governmental departments and agencies have a huge opportunity to capitalize on data, AI, and other emerging technologies to drive innovation to achieve mission outcomes and deliver at scale and market speed. They will be empowered to provide a better and more personalized experience for citizens, optimize processes, reduce time on basic tasks to focus on evidence-based decision making and innovative ways to improve services.
The NTIS framework provides a unique framework to achieve these objectives in a rapid, efficient, and effective way.
References
Action Plan — Federal Data Strategy. (2020). Retrieved from Strategy.data.gov:
https://strategy.data.gov/action-plan/
Artificial Intelligence Initiative Act. (2019–2020). Retrieved from Congress.gov: https://www.congress.gov/bill/116th-congress/senate-bill/1558/all-info
Executive Order on Maintaining American Leadership in Artificial Intelligence. (2019). Retrieved from whitehouse.gov: https://www.whitehouse.gov/presidential-actions/executive-order-maintaining-american-leadership-artificial-intelligence/
The Paper Reduction Act. (1995). Retrieved from govinfo.gov:
https://www.govinfo.gov/content/pkg/PLAW-104publ13/pdf/PLAW-104publ13.pdf
Is the Federal Government Ready for AI? (2019). Retrieved from Govexec.com: https://www.govexec.com/insights/federal-government-ready-ai/
The AI hierarchy of needs. (2017). Retrieved from Hacker Noon:
https://hackernoon.com/the-ai-hierarchy-of-needs-18f111fcc007
Request for Comments on a Draft Memorandum to the Heads of Executive Departments and Agencies, ‘‘Guidance for Regulation of Artificial Intelligence Applications’’. (2020). Retrieved from govinfo.gov: https://www.govinfo.gov/content/pkg/FR-2020-01-13/pdf/2020-00261.pdf
U.S. Leadership In AI: A Plan for Federal Engagement in Developing Technical Standards and Related Tools. (2019). Retrieved from nist.gov: https://www.nist.gov/system/files/documents/2019/08/10/ai_standards_fedengagement_plan_9aug2019.pdf
