Skip to main content

How the US government can accelerate AI entrepreneurship

The US government can play a big role in helping AI entrepreneurs and, thus, accelerating AI innovation

Artificial Intelligence (AI) may be the most impactful technology of our generation. From how we find and consume media, communicate with others, and move about in our environments, it already affects several aspects of human life and increasingly catalyzes the global economy.

For these reasons, the US government has made AI innovation a top priority. Congress has passed major legislation to support AI development and adoption, including the 2020 National AI Initiative Act and the 2020 AI in Government Act.1 In 2021, the National Security Commission on AI recommended doubling nondefense funding for AI research and development annually to reach US$32 billion per year by 2026.2

A primary source of tech innovation has traditionally been entrepreneurs and small businesses. Small firms regularly produce more patents per employee than large firms, and these patents are also associated with higher impact, growth, and profits.3 Additionally, small businesses account for 44% of US gross domestic product (GDP) and create two-thirds of net new jobs.4

Advancing AI innovation in the United States will therefore likely require greater support and opportunity for AI entrepreneurs and small businesses. Unfortunately, they currently confront several challenges to bringing their innovations to market and finding a more sustainable business footing. The US government can play essential roles to help new and emerging AI firms and overcome the challenges that can hinder their success, thereby supporting its broader aim of accelerating AI innovation. In three distinct roles—as buyer, regulator, and infrastructure provider—the government can use its influence to promote an environment in which emerging AI technology firms can thrive.

This paper begins with a brief overview of the potential benefits of and challenges to AI innovation. We then highlight some of the approaches the US government can apply in each of its three roles to help new AI entrepreneurs and startups in commercial and government markets. Such efforts can help advance new innovations to ensure society reaps the full range of AI benefits and the United States remains a world leader in this technological revolution.

Attractive benefits, but formidable challenges

Artificial intelligence encompasses an array of statistical, machine learning, and computational technologies that enable systems to learn, respond, make decisions, and take actions with increasing autonomy. In practice, AI is central to emerging technologies such as autonomous vehicles, virtual assistants, and ubiquitous recommendation engines found across entertainment services, online retailers, and social media, among many other applications.

These AI capabilities promise to unlock new productivity gains and innovations for workers, private businesses, and government. Research estimates that AI can contribute 14.5% to North American GDP, equivalent to US$3.7 trillion, by the year 2030.5 The use of AI can also improve the accuracy and precision of service delivery by government agencies, saving up to 1.3 billion employee hours annually according to analysis by Deloitte.6 Lastly, as both a cutting-edge technology and a large industry, AI has substantial implications for national security. The US National Security Commission for AI concluded in 2021, “No comfortable historical reference captures the impact of artificial intelligence (AI) on national security.”7

Several challenges make capturing the full benefits of AI difficult, however. This is particularly true for the unique innovations that often emerge from new and emerging technology firms. These challenges include the increasing cost of computing resources, a lack of high-quality data, and a shortage of AI talent. To be sure, all companies face these challenges, but they can be existential for small businesses trying to cross the so-called “valley of death” in which new firms must survive an extended period of negative cash flow before generating positive income.8

  • High costs of computing. Developing a commercialized AI application is an iterative process—models need to be consistently retrained to improve performance and retain model accuracy. Every iteration of model training typically requires significant resources, which can result in a very expensive process. Increasingly, the cost of computational resources is a major component of this expense. Between 2012 and 2018, for instance, premier AI algorithms doubled their computational power consumption every 3.4 months.9 In 2019, one AI executive shared his belief that AI experimentation will soon “hit a wall” due to the costs of computing power.10 These costs risk crowding out would-be AI entrepreneurs and small businesses.
  • Low-quality data. Models require vast amounts of high-quality data to reach the performance and robustness thresholds that make them commercially viable. However, 80-90% of available data is unstructured, which needs to be “cleaned” to be ingested by models for processing.11 As illustrated in the chart below from Anaconda’s 2020 State of data science report, cleaning data is the most time-intensive phase of model development, taking about 26% of a data scientist’s time.12 The time, and money, required to clean data can be costly; running out of capital is the number one reason why emerging technology firms fail.13
  • Scarcity of talent. AI application development calls for unique talent. There is a significant dearth of technical talent such as data scientists, solution architects, software developers, data strategists, and DevOps professionals in the United States. It is estimated that 250,000 data scientist jobs remain unfilled.14 By 2030, the shortage of all technical talent in the United States could reach 6 million people, a figure that could cost US$162 billion in revenue.15 This can be especially challenging for small businesses, which have to compete with large tech companies that tend to have better brand recognition, compensation and benefits, and greater job security.16
  • Hurdles to government contracting. Government contracting also has important and unique impacts on small tech firms. Some research has found that technology startups produce more disruptive and impactful innovations when leveraging public funds—whether in the form of grants, cooperative R&D agreements, or contracts—than private funding.17 Through direct funding for high-risk, high-reward early research, such as the Internet, or by using federal contracts to build the initial market and fast-track the process to commercialization, such as microchips, the US government has made and continues to make a significant impact on technology innovations.18

To protect national and cybersecurity, the United States has implemented several processes and procedures that federal agencies must follow to procure private sector technology, including the Federal Risk and Authorization Management Program (FedRAMP) and Authorization to Operate (ATO) processes.19 Although warranted, these pose challenges to using commercial AI technology in government given the high cost and lengthy processing time to certification, challenges felt more acutely by new and emerging entrepreneurs. In fact, startups trying to pursue government contracts are often discouraged by the “labyrinthine” contracting compliance process they face, which can include years of paperwork and hundreds of thousands of dollars in fees for legal and contracting expertise, all of which create significant challenges to entry.20 Surveys of technology startups indicate that the amount of time it takes to secure government contracts and the complex and demanding proposal process are the two leading factors discouraging them from submitting bids.21

These inefficiencies translate into reduced opportunities for AI entrepreneurs and, potentially, less AI innovation from startups. There has been some recent headway to simplify contracting processes and requirements, but the bottom line is clear: Agencies should continually seek to improve the federal technology procurement process so that the US government can pilot, test, commercialize, and reap the innovation benefits of AI entrepreneurs.22

Government’s roles in spurring innovation

While the issues identified above can create significant challenges to emerging technology firms, there are many approaches government can use to overcome these challenges and enable AI innovation from small businesses. Most can be grouped into three roles that the government can pursue:

  • Directly fund AI innovation (as a buyer)
  • Write, update, and pass legislation that streamlines existing policies and regulations and creates new incentives for AI entrepreneurship (as a regulator)
  • Provide critical technological infrastructure required for advanced AI development (as an infrastructure provider)

The accompanying figure categorizes all these approaches by their respective government roles and how they can help address the challenges facing AI entrepreneurs:

Buyer

Increased government spending on pilot programs

Pilot programs provide an opportunity for smaller firms to navigate government contracting and for government agencies to identify how they can streamline their processes for AI entrepreneurs. A US$600 million pilot from the Department of Defense (DoD) to test and experiment with 5G technology at five military installations provides one model of how to create a sizable opportunity for smaller firms focused on cutting-edge and sensitive technologies.23 The long-term impact of these programs, however, is contingent on a transition to programs of record or contract vehicles with sustained and predictable funding.24

Government venture capital

The US government can build on the high-risk, high-reward approach that venture capital investing takes to support new AI startups. Initiatives such as In-Q-Tel25 at the Central Intelligence Agency, AFVentures26 at the Air Force, and a partnership between Health and Human Services and Global Health Investment Corporation27 demonstrate how agencies can engage the private sector to support agency missions and bring innovations to market.

Agency-university direct partnerships

Beginning in 2020, the National Science Foundation launched its AI Research Institutes initiative. This program has since awarded US$360 million to 18 universities across 40 states.28 A critical element of the initiative is a focus not merely on research and development but on the translation of these results into market and practice and on workforce networking.29 This ecosystem approach can deliver AI innovations and bring AI entrepreneurs to the marketplace.

Government funding for advanced AI research

Federal agencies requested a total of US$1.7 billion for AI research and development in nondefense areas in fiscal year 2022.30 The National Security Commission for AI recommended spending US$32 billion annually in nondefense-related AI research and development by 2026; so there is extensive room for growth.31

Regulator

Private sector tax incentives

The US tax code enables the government to support specific sectors and entities by making modifications to reduce tax liabilities. To incentivize greater sharing of data and compute resources from larger to smaller AI enterprises the government could explore tax exemptions and reductions for technology transfer. Additionally, compliance requirements could be simplified. The complexity of applying for the R&D tax credit can crowd out smaller firms, which claim less than 10% of such credits.32 The standard R&D tax credit is 20% of qualifying expenditures, but such credits could be increased for small businesses.

Data flow protections, domestically and internationally

The US government could prioritize the completion of Office of Management and Budget (OMB) guidance for agencies on how to adhere to the Open, Public, Electronic and Necessary (OPEN) Government Data Act so that more high-quality data is made available for AI innovation.33 The US government could pursue bilateral and multilateral efforts to improve cross-border data-sharing such as relevant provisions added to the United States-Mexico-Canada Agreement (USMCA).

Infrastructure provider

Enable regional innovation hubs

A few coastal technology hubs (Boston, San Francisco, San Jose, Seattle, San Diego, etc.) have accounted for more than 90% of the growth in the high-technology sector.34 To tackle these issues, the US government can drive more investment to historically nontechnical metro areas into AI innovation hubs by building on programs such as the Opportunity Zone and the SBA Historically Underutilized Business (HUB) zones.35

The US government can also establish a competitive process in which metro areas apply and present their business case to be chosen as an innovation hub and receive funds to underwrite relevant activities, such as supporting higher education institutions or building infrastructure to attract talent and build tech districts. Applications from tech-limited metro areas or often overlooked rural regions could be privileged to ensure funds are invested in underserved areas. Additionally, such efforts may help address low representation of minority groups in AI talent and small businesses.

How the US government can help

In addition to the wide variety of general approaches discussed above, there are a few specific potentially impactful approaches that address computing costs, low-quality data, scarcity of talent, and contracting challenges:

Provide access to low-cost computing infrastructure

The largest technology firms in the United States (Google, Amazon, Facebook, Apple, and Microsoft)36 have access to a wide breadth of computing infrastructure. Three of these firms (Amazon, Microsoft, and Google) account for about 62% of the global cloud infrastructure market.37 These firms are also prominent in the government cloud market, as they are among the top 10 in terms of the number of FedRAMP authorizations received, with Amazon topping the list.38

The US government can incentivize large technology firms to share their computing infrastructure with emerging technology firms and to build upon their existing FedRAMP authorized services. For example, tax breaks can include an accelerated schedule for the depreciation of physical R&D assets (mainframes, servers, processors, etc.). This can minimize taxable income and allow for greater tax deductions in the earlier years of an asset, thereby encouraging more regular investments and upgrades to physical R&D assets. Emerging technology firms can focus their capital on application development, rather than infrastructure and the FedRAMP authorization process, accelerating timelines to application commercialization and access to the government marketplace.

For nonprofit institutions, the US government can build on current programs such as the National Science Foundation–funded initiative to allow for easy access to their supercomputers.39 The government can incentivize institutions such as the DOE’s 17 national laboratories or US universities to provide emerging technology firms low-cost access to their computing infrastructure, initiatives that might be supported by the National Artificial Intelligence Research Resource (NAIRR) as it is stood up in 2023. Incentives can include government grants or tax incentives for universities to donate computing infrastructure that has been ear-marked for upgrading. More generally, the NAIRR could provide a coordinating hub to help new AI entrepreneurs and startups identify available low-cost computing power. Like the FedRAMP example, this nonprofit institution approach could accelerate commercialization timelines for emerging technology firms.

Contracting instruments and programs

The government has developed several contracting instruments and programs to improve technology acquisition. The first program is Tradewind, an AI-focused acquisition platform for the Department of Defense (DoD) that works with government, industry, and academia to accelerate AI adoption. The DoD’s Joint Artificial Intelligence Center (JAIC) developed Tradewind to revitalize its procurement process for AI solutions. It aims to provide JAIC experts access to emerging technology firms and academia while enabling the DoD to award prototype contracts in less than two months and operationalize critical AI technologies.40

Another DoD organization, the Defense Innovation Unit (DIU), awards scalable contracts to commercial technology companies focused on national security challenges. The DIU helps DoD entities identify technology gaps and matches them with technology companies that provide solutions, reducing costs for companies aiming to partner with the DoD.41

Outside of the DoD, there are programs aimed specifically at small businesses such as the Small Business Innovation Research Program (SBIR) and the Small Business Administration (SBA) Mentor-Protégé Program. The SBIR is a competitive research program that enables US small businesses in the technology sector to engage in federal research and development while providing incentives to profit from commercialization.42 The SBA Mentor-Protégé program enables small businesses to gain business development skills and win government contracts through partnerships with more experienced companies.43

These four instruments and programs provide templates and leading practices that the US government can leverage while using government funding to streamline the commercialization of emerging technology.

Show more

Convene a multiparticipant cross-sector consortium

To address the issue of low-quality data, the US government can convene a multiparticipant cross-sector consortium to enable the ecosystem for AI application development. For instance, the European Union (EU) Partnership on AI, Data and Robotics is a consortium of government, industry, academia, and nonprofit leaders. It aims to provide the ecosystem necessary to guide early-stage AI research, right from the idea stage to fully deployed applications. This ecosystem includes stimulating industrial investments, boosting academia-industry collaborations, and mobilizing the community to create horizontal and vertical partnerships.

An example of a cross-sector consortium in the United States is the First Five Consortium, cochaired by the DOE’s Artificial Intelligence and Technology Office (AITO) and Microsoft. It was created in August 2020 to develop AI solutions related to humanitarian assistance and disaster response.44 The First Five Consortium’s scope is more limited than that of the EU’s consortium, which applies to all use cases in the connected fields of AI, data, and robotics. The US government can build on this consortium, using the EU Partnership as a template to create a consortium with a broader mandate in covering AI-related topics in the United States. It can leverage the capabilities of world-class American universities, technology firms, and nonprofits to drive domestic AI innovation by addressing shortages in high-quality data and available computing resources.

Develop open-market work visas for foreign AI talent

To address scarcity of talent, the US government can build on current immigration laws to allow foreigners with AI skills to live and work in the country without being tied to a sponsoring employer. The current visa programs for high-skilled nonimmigrants—the H1-B visa and the Optional Practical Training (OPT) component of student F-1 visas—are limited by a hard cap on visas per year and sponsoring employer approval requirements.45

Instead of sponsorship, applicants could be required to work in an AI-related field or firm for a lengthy period—potentially 10 years—before becoming eligible to apply for a green card or naturalization. This would level the playing field for emerging technology firms, which traditionally do not have the resources to sponsor immigrants, thus helping them attract and retain top international AI talent. Rigorous upfront screening of student visa applications by the US government could prevent security concerns, while future opportunities to become fully naturalized can incentivize applicant adherence to the process.46 Together these adjustments could expand the ways for top talent to work in the United States, alleviating the shortage of available workers with the highly specialized skill sets needed in the AI sector.

Reform and build on existing contracting programs

Reforms are currently being discussed to improve the FedRAMP program to speed up the compliance and certification process for small businesses.47

Among the aims of the reforms are:

  1. Allocation of additional annual funding to the program
  2. Reduction of duplication of security assessments by promoting agencies to use cloud technologies that have been already certified by another agency
  3. Tasking the US General Services Administration to work toward making the program more efficient by automating security assessments and continuous monitoring processes.48

Additionally, the FedRAMP Agency Liaison Program was recently established to revamp how the FedRAMP program communicates and works with federal agencies. The Liaison Program employs a “train-the-trainer” model with the goal of empowering agency liaisons to distill information about the authorization process to make it more efficient.49 To create impact, the US government can build on the FedRAMP Agency Liaison Program by establishing agency-specific AI application testing and certification environments that partner with emerging technology firms to accelerate their pathway to FedRAMP status. In this proposed partnership, agency liaisons can work with emerging technology firms as they build and test their AI applications so that the firms can build AI applications that adhere to FedRAMP requirements and gain access to government computing resources to commercialize their applications. These new resources can help lower the compliance challenges new AI entrepreneurs often confront.

Final considerations

Artificial intelligence is poised to have dramatic impacts across the global economy. However, a range of challenges currently complicate the work of entrepreneurs and small businesses that could be the sources of much AI innovation. There is much that the US government can do—as a buyer, regulator, or infrastructure provider—to overcome these challenges, thereby accelerating AI entrepreneurship and economic growth. By applying the right capabilities with sufficient urgency, the government can foster innovation and promote an environment in which emerging AI technology firms thrive.

While government efforts to accelerate AI innovation and entrepreneurship are critical, success in these areas will raise new challenges around ethical and reliable AI. While we lack available time and space to adequately address these topics here, we want to close by emphasizing that guardrails relating to equity and trustworthiness remain critical complements to AI innovation. The below considerations are a starting point—in practice, each requires a nuanced execution strategy to be successful. Specifically, the US government should:

  • Make sure diverse perspectives are heard and discussed in all public-private partnerships, particularly the National Artificial Intelligence Advisory Committee at the Department of Commerce and other task forces, interagency working groups, and other forums discussing US policy toward AI.
  • Establish data governance practices to mitigate potential cybersecurity and national security challenges consistent with the 2030 Vision of the Federal Data Strategy and the May 2021 Executive Order on Improving the Nation’s Cybersecurity.
  • Build on the numerous guidelines that federal agencies are developing to support trustworthy and ethical use of AI. In particular, the National Institute for Standards and Technology is synthesizing an AI risk management framework to be released in 2023. The starting point for AI initiatives at Deloitte is its own Trustworthy AI framework.50

A dual focus on AI innovation and ethical AI should be mutually reinforcing, fueling technological gains and benefits that are widely shared and supported in the United States and around the world.

About the Deloitte AI Institute

The Deloitte AI Institute helps organizations connect all the different dimensions of the robust, highly dynamic, and rapidly evolving AI ecosystem. The AI Institute leads conversations on applied AI innovation across industries, using cutting-edge insights to promote human-machine collaboration in the Age of With™. The Deloitte AI Institute aims to promote dialogue about and development of artificial intelligence, stimulate innovation, and examine challenges to AI implementation and ways to address them. The AI Institute collaborates with an ecosystem composed of academic research groups, start-ups, entrepreneurs, innovators, mature AI product leaders, and AI visionaries to explore key areas of artificial intelligence including risks, policies, ethics, future of work and talent, and applied AI use cases. Combined with Deloitte’s deep knowledge and experience in artificial intelligence applications, the institute helps make sense of this complex ecosystem and, as a result, delivers impactful perspectives to help organizations succeed by making informed AI decisions. No matter what stage of the AI journey you’re in—whether you’re a board member or C-suite leader driving strategy for your organization, or a hands-on data scientist bringing an AI strategy to life—the Deloitte AI institute can help you learn more about how enterprises across the world are leveraging AI for a competitive advantage. Visit us at the Deloitte AI Institute for the full body of our work, subscribe to our podcasts and newsletter, and join us at our meetups and live events. Let’s explore the future of AI together.

Learn more

  1. Congress.gov, “H.R.2575 - AI in Government Act of 2020,” September 14, 2020.

    View in Article
  2. National Security Commission on Artificial Intelligence, The final report, accessed August 18, 2022.

    View in Article
  3. Anthony Breitzman and Diana Hicks, “An analysis of small business patents by industry and firm size,” Faculty Scholarship for the College of Science & Mathematics, November 1, 2008; William Baumol, “Small firms: Why market-driven innovation can’t get along without them,” in The Small Business Economy 2005: A Report to the President (Washington, DC: US Government Printing Office, 2005), pp. 183–206.

    View in Article
  4. US Small Business Administration Office of Advocacy, “Small businesses generate 44 percent of U.S. economic activity,” press release, January 30, 2019.

    View in Article
  5. PricewaterhouseCoopers, Sizing the price: What’s the real value of AI for your business and how can you capitalise?, accessed August 18, 2022.

    View in Article
  6. Peter Viechnicki, William Eggers, How much time and money can AI save government?, Deloitte Insights, April 26, 2017. 

    View in Article
  7. NSCAI, The final report.

    View in Article
  8. Martin Zwilling, “10 ways for startups to survive the valley of death,” Forbes, February 18, 2013.

    View in Article
  9. Will Knight, “Prepare for artificial intelligence to produce less wizardry,” Wired, accessed January 19, 2022.

    View in Article
  10. Will Knight, “Facebook’s head of AI says the field will soon ‘hit the wall,’Wired, accessed January 19, 2022.

    View in Article
  11. Tam Harbert, “Tapping the power of unstructured data,” MIT Sloan School of Management, February 1, 2021.

    View in Article
  12. Anaconda, 2020 State of data science, June 30, 2020.

    View in Article
  13. CB Insights Research, “The top 12 reasons startups fail,” August 3, 2021.

    View in Article
  14. Jen DuBois, “The data scientist shortage in 2020,” QuantHub, April 7, 2020.

    View in Article
  15. Michael Franzino et al., “The $8.5 trillion talent shortage,” Korn Ferry, July 26, 2021.

    View in Article
  16. Daisy Wright, “Why job seekers may choose larger firms over small businesses,” The Balance Careers, February 9, 2021.

    View in Article
  17. Luke Sykora, “Is government funding good for your startup?,” Stanford eCorner, July 3, 2019.

    View in Article
  18. Peter L. Singer, Federally supported innovations: 22 examples of major technology advances that stem from federal research support, The Information Technology & Innovation Foundation, February 2014.

    View in Article
  19. FedRAMP website

    View in Article
  20. Daniel K. Lim, “Startups and the Defense Department’s compliance labyrinth,” War on the Rocks, January 3, 2022; Shands Pickett, “Reboot the defense innovation ecosystem,” War on the Rocks, April 7, 2022.

    View in Article
  21. Geoff Orazem et al., “Why startups don’t bid on government contracts,” Boston Consulting Group, August 22, 2017.

    View in Article
  22. Xuli Tang et al., “The pace of artificial intelligence innovations: Speed, talent, and trial-and-error,” arXiv.org, September 3, 2020.

    View in Article
  23. US Department of Defense, “DOD announces $600 million for 5G experimentation and testing at five installations,” news release, October 8, 2020.

    View in Article
  24. Joe Gould, “New Pentagon strategy to bring in small businesses coming soon,” Defense News, March 7, 2022.

    View in Article
  25. Damian Paletta, “The CIA’s venture-capital firm, like its sponsor, operates in the shadows,” Wall Street Journal, August 30, 2016.

    View in Article
  26. Mark Sullivan, “Why the Air Force has its own venture capital fund,” Fast Company, April 6, 2021.

    View in Article
  27. Heather Landi, “HHS launches $50M venture capital partnership to develop tech for pandemic response,” Fierce Healthcare, June 1, 2021.

    View in Article
  28. Brandi Vincent, “NSF adds 11 new AI research institutes to its collaborative, nationwide network,” Nextgov, August 4, 2021.

    View in Article
  29. Jason Bates, “Expanding the geography of innovation: NSF AI Research Institutes 2021,” National Science Foundation, July 29, 2021.

    View in Article
  30. National Science and Technology Council, The Networking & Information Technology R&D Program and the National Artificial Intelligence Initiative Office Supplement to the President’s FY2022 Budget, December 2021.

    View in Article
  31. NSCAI, The final report.

    View in Article
  32. Alex Muresianu and Garrett Watson, “Reviewing the federal tax treatment of research & development expenses,” April 13, 2021.

    View in Article
  33. US Government Accountability Office, “Open data: Agencies need guidance to establish comprehensive data inventories; information on their progress is limited,” October 2020.

    View in Article
  34. Brookings Institution and Information Technology & Innovation Foundation, The case for growth centers: How to spread tech innovation across America, December 2019.

    View in Article
  35. Internal Revenue Service, “Opportunity Zones frequently asked questions,” December 15, 2020; US Small Business Administration, “HUBZone program,” accessed August 18, 2022.

    View in Article
  36. This article is an independent publication and has not been authorized, sponsored, or otherwise approved by Apple Inc.

    View in Article
  37. Synergy Research Group, “Cloud market ends 2020 on a high while Microsoft continues to gain ground on Amazon,” February 2, 2021.

    View in Article
  38. FedRAMP, “The Federal Risk and Management Program dashboard,” accessed August 18, 2022.

    View in Article
  39. Ken Chiacchia, “PSC and partners to lead $7.5-million project to allocate access on NSF supercomputers,” Pittsburgh Supercomputing Center, April 22, 2022.

    View in Article
  40. Tradewind website

    View in Article
  41. Defense Innovation Unit website

    View in Article
  42. SBIR.gov, “The SBIR and STTR programs,” accessed August 18, 2022.

    View in Article
  43. US Small Business Administration, “SBA Mentor-Protégé program,” accessed August 18, 2022.

    View in Article
  44. US Department of Energy, “Department of Energy announces the First Five Consortium,” August 18, 2020.

    View in Article
  45. USCIS, “H-1B specialty occupations, DOD cooperative research and development project workers, and fashion models,” February 5, 2021.

    View in Article
  46. National Foundation for American Policy, “H-1B visas by the numbers: 2017-18,” April 2018; Martijn Rasser, The American AI century: A blueprint for action,” Center for a New American Security, December 17, 2019.

    View in Article
  47. FedRAMP, A case study for change within government: Reflections from spring 2017, accessed August 18, 2022.

    View in Article
  48. Congress.gov, “H.R.21 - FedRAMP Authorization Act,” January 6, 2021.

    View in Article
  49. FedRAMP, “FedRAMP announces Agency Liaison Program,” June 24, 2020.

    View in Article
  50. Tasha Austin and Ed Van Buren, “Developing and deploying trustworthy AI in government,” Deloitte, August 2021.

    View in Article

The authors would like to thank Davin O’Regan and Joe Mariani for their thoughtful feedback on the drafts.

Did you find this useful?

Thanks for your feedback

If you would like to help improve Deloitte.com further, please complete a 3-minute survey