A MAJOR TWO-YEAR STUDY CONDUCTED by a newly formed commission on AI opened its report with these words. “Americans have not yet grappled with just how profoundly the artificial intelligence (AI) revolution will impact our economy, national security, and welfare.” This statement referred to the American masses. One could ask that despite AI being talked about in magazines, blogs, news sites, and popular media, how is it possible that America has not grappled with the effects of the AI revolution? The answer lies in developing an appreciation of how societies make sense of technological changes. Being fascinated with technology is one thing, adopting it another. Being able to adopt a technology project is one thing, approaching technology-based transformation another. Experimenting with technology is something else, strategic diffusion something totally different. The commission's report was pointing to the fact that even though Americans were in general familiar with AI, they were missing the depth and breadth of the impact the technology will have on them, their country, their economic well-being, and their lives. This assessment was not only true for the masses, but it was also true for the American leaders.

Apparently, sensing that OSTP's whitewashing and self-praising passive approach to AI was creating an existential competitive threat to the US, Congress and the executive branch sponsored a different commission to study AI from a national perspective.

BARBARIANS AT THE GATE

By 2021, AI was no longer a new idea for America. Massive investment was going into the technology, the start-up scene was getting hotter by the day, and companies had begun to talk about AI. In 2019, and then in 2020, OSTP had already declared victory. If it were a scorecard, OSTP self-awarded straight As to itself. The eight strategies were considered necessary and sufficient, a largely superficial or opportunistic feedback mechanism had been architected, and there was no corresponding performance measurement system to measure the progress made. There was no critical thinking. There was no avenue for reflection. What was being measured and claimed as performance were the activities and not the results. OSTP patted its back in terms of introducing AI in American government, conducting a few conferences, and issuing a couple of reports that highlighted the eight strategies. There was no sense of urgency, no anxiety, no national inspiration, no Sputnik moment, no information superhighway–like mobilization. The strategy was as boring as a grocery shopping list.

While OSTP stayed in its bubble, greatly fascinated with its own eight strategies, another AI study was launched by the government. This one was sponsored jointly by Congress and the executive branch. As the results of the study were published, they painted a completely different picture from the rosy picture given by the OSTP. The bottom line of the finding: America was not ready for the AI competition.

Notice that the finding was not that America was lagging in the competition or losing to a worthy adversary. The finding was much more problematic: America was not even ready to play the game. The former interpretation would be analogous to an athlete who qualifies for the Olympics, participates in the events, makes it to the final, and then as the race draws to the final round, for one reason or another falls behind a better competitor. The second, and far more problematic, is where the athlete would not even qualify to be in the Olympics because the athlete was not prepared to do so. The warning was dire—but neither did it have any major impact on national leadership nor was it properly understood.

America seemed to be in a competitive disadvantage, and the report authors did not hide or cover up the findings. This project lasted over two years, and at the end a powerful report was issued.

THE FINAL REPORT

Recognizing an urgent need to perform a thorough investigation, Congress and the executive branch converged their forces to analyze and develop a realistic picture of artificial intelligence in America. A special commission was established to investigate AI in America. More than 700 pages of draft materials were posted for public review and comment, and 15 hours of deliberations were streamed online. The commission known as National Security Council of Artificial Intelligence was established as a Federal Advisory Committee. Headed by Google's former CEO Eric Schmidt (chairman) and Robert Work (vice chairman), the commission investigated and then issued a report. We start by first identifying some of the findings.

The report began by outlining its conclusions:

Despite our private-sector and university leadership in AI, the United States remains unprepared for the coming era. Americans must recognize the assertive role that the government will have to play in ensuring the United States wins this innovation competition. Congress and the President will have to support the scale of public resources required to achieve it.

The magnitude of the technological opportunity coincides with a moment of strategic vulnerability. China is a competitor possessing the might, talent, and ambition to challenge America's technological leadership, military superiority, and its broader position in the world. AI is deepening the threat posed by cyber attacks and disinformation campaigns that Russia, China, and other state and non-state actors are using to infiltrate our society, steal our data, and interfere in our democracy. The limited uses of AI-enabled attacks to date are the tip of the iceberg. (NSCAI 2021)

The above excerpt describes the problems that the commission observed. They acknowledged that America does have private-sector and university leadership in AI, and yet America remains unprepared. One might ask, how is it possible that America has private-sector and academic and research leadership and yet “remains unprepared”? Aren't these two—private-sector capability and academic research potential—the essential foundations of progress? What else would one need? Where are the cracks? This was the first time that the OSTP-led field of dreams was being exposed. This was where five years of self-fascination was being challenged in a more objective manner. A sense of reality was emerging. It was also recognized that this naïve thinking that somehow public-private partnership would automatically result in progress was not delivering results. Clearly, preparing for the competition meant a lot more than just the two parties—academia and business—possessing talent. America was unprepared.

The second sentence clarified “the assertive role that the government will have to play.” What does that exactly mean? Does it mean that the commission is asking the government to play a more active and invasive role? The commission was clarifying that the role played by the government was not good enough and that it lacked what needs to get done to drive true results. America was falling behind due to the passiveness and lack of leadership from the government. An assertive role is a role of leadership with clarity. It is not biased, and it is not self-serving. The commission added that China is not just any competitor. It is a competitor that possesses “might, talent, and ambition” to challenge America's technological leadership, military superiority, and its broader position in the world. This was an important sentence as it described the nature of the competition being technological leadership, military superiority, and the broader position in the world—which includes economy, society, soft power, values such as democracy, and institutions. The competitor had the resources and the motivation to unseat America. Later on in the report, the commission pointed out that “China is organized, resourced, and determined to win this contest.” This was a powerful insight to bring attributes of competition into the picture. Unlike the OSTP planning, which lacked the competitive need and the urgency to drive results, the commission introduced a much-needed new perspective: America is not alone, America cannot be complacent, and America must work at the speed of relevance.

Lastly, the report attributed new types of risks to the nation that were emerging from the AI-related supremacy of the adversaries. Thus, not only a loss in AI competition will reduce and diminish the power of America, but it will also actually accelerate and intensify the decline. AI is not only the best defense, it is also the best offense.

WHAT WAS HUSHED

While the report pointed out America's shortcomings and opportunities missed, the commission did not dwell into the root causes or drivers for those problems. That position is understandable.

The first was a hint toward the problems—some of which we have pointed out in the previous chapters. The commission reports, “While a competitive response is complicated by deep academic and commercial interconnections, the United States must do what it takes to retain its innovation leadership and position in the world” (NSCAI 2021, 11). We interpret these words to gently point to the nepotism and symbiotic relationship that exists in the tech sector and that destroys value for the country. But after stating that, the commission's report quickly moves forward to sidestep the issue by saying that the US must do whatever it takes. The above words in the commission's report are extremely important, and they should not be ignored. The commission says that a competitive response will be complicated due to deep academic and commercial interconnections and then adds that the US “must do what it takes” without suggesting what those actions should be. For a start, we strongly recommend that policy should not be developed by conflicted professors and Big Tech. We also recommend that this is not time to go easy on what's broken and who's responsible. The time to fix things is slipping fast. The commission acknowledges that but also hedges—most likely to not offend others. We don't think that should be a concern. This is an existential threat for national competitiveness, and it is important to be open and assertive about it.

The second part where the commission hedged was when the report acknowledged that previous warnings were ignored and that new warnings are now issued and if these are not heard, the United States will lose its leadership position. “This commission can warn of national security challenges and articulate the benefits, rather than explain why previous warnings were ignored and opportunities were missed” (NSCAI 2021, p. 19). The problem with this position is that while the report accepts the assertions that “the AI revolution was not a strategic surprise” and that previously issued guidance was ignored, it does not bother to dwell into the reasons why that happened. Most likely that was done for political expediency. Since commercial interests now dominate national interests, this has become the DC way of handling problems: identify the problem and then insist upon moving on without exploring the root causes.

This position typically ignores the root causes and the related underlying processes. It is easier to move the goalpost than to expose the performance gaps of someone. Again, the commission concludes that there were warnings that were ignored. And now there is a new warning issued by the commission, and a new guidance is provided. How can they be sure that the new warnings will not be ignored? How do they even know that the previous warnings were ignored? Perhaps they were not ignored, but the way work was done was clumsy and the execution was flawed? Perhaps the previous warnings were not understood or communicated properly? Perhaps the vision was poorly defined? Perhaps we failed to inspire people? Perhaps we lacked the managerial capacity? Perhaps the underlying organizational model, property rights, or the economic structure needed revision? Perhaps the strategy was not properly developed, articulated, or communicated? The commission accepted the blinders and refused to look deeper into the problem.

We believe that exposing and understanding the root causes was important. Unless we change those underlying conditions, every plan and goalpost move will have the same result. No warning can change things when the entire structure is designed to become an echo chamber of complacency, deception, political bickering, and exaggerations. The report refused to confront these issues.

We have taken a completely different position in this book. We are identifying the problems so that sincere and clear strategy can emerge from true leadership. We are not hiding or covering the problems. We are not hedging or evading.

The commission declared that:

Despite exciting experimentation and a few small AI programs, the U.S. government is a long way from being “AI-ready.” The Commission's business leaders are most frustrated by slow government progress because they know it's possible for large institutions to adopt AI. AI integration is hard in any sector—and the national security arena poses some unique challenges. (NSCAI 2021, p. 2)

This was the reality that America must face. The five years of the OSTP strategy had delivered the following: exciting experimentation and a few small programs. Not just that, the US government was far from being AI-ready—note the reference was not to AI-capable or AI-leader or AI-expert but to the basic requirement of being AI-ready. In other words, we were not even qualified to enter the arena. This was in stark contrast to what OSTP was feeding the two presidents and to America. The OSTP was claiming that America was at the forefront, prepared, a leader in all respects, and at the cutting edge. See Figure 6.2, and you would think that all agencies have now implemented long-term AI plans and that such plans are showing great results. After all, the majority of the boxes were checkmarked. This shows the cognitive dissonance of the OSTP.

Despite the two shortcomings of the commission's report—the reluctance to directly challenge the problems of the complex relationship between academia and commercial interests and the hesitation to explore and expose the root cause analysis of America's failure to rise to the top—the report gives a reasonable and accurate representation of the state of AI in America. We agree with the report's representation; however, there are five areas that the report did not cover, and our discussion below will complement the findings and the recommendations of the report.

1. MORE MONEY

The finding articulates that the buck stops at the president and Congress and recommends supporting the scale of public resources. Our approach to funding is different. While we agree with the commission to increase funding, we believe it is extremely important to develop an investment allocation model that uses multidimensional criteria to allocate money. Just as the OSTP did, the commission was also viewing the solution to the problem as throwing more money into AI, and everything would take care of itself. This represents the interests of the various research groups, Big Tech, which views this as subsidizing their research, and associations but is not necessarily in the best interest of the country. A plan for industrialization of AI must exist before money is thrown into the pit. The strategic orientation of investment must precede the funding. We must avoid the short-term, opportunistic mindset of OSTP: build it, and they will come. In Chapter 12 we show what would constitute a comprehensive plan.

We also disagree with this approach as this puts the cart before the horse. This recommendation style is symptomatic of the problem: everything is being approached from a short-termism. This also undermines what is the real role and responsibility of the president and Congress. Both should not be viewed solely as check writers. We must expect more from our leaders whose conduct, thinking, vision, communication, and leadership provides the inspiration that moves nations forward. The fundamental mistake of viewing Congress and the president as some bankers should not be repeated. Current and future presidents and members of Congress should receive comprehensive training on AI and the related economic and other drivers.

This spirit of just throwing in more money can work in less revolutionary times and when the competitive distance is wide. But when the competitive difference is short and technological lead has diminished, it is important to accelerate innovation in targeted domains first. This can stop the bleeding and give some quick wins. Innovations embed information and find their own paths and trajectories as they move through their life cycle. Information gets embedded at every level in the progression. The paths are adjusted and then readjusted. To bring greater certainty into innovation, one must incur more costs related to information. One must collect information about the future direction of the technology and embed that in the research and development process. When you are falling behind, you need to invest in getting greater information and not wait for technologies to fight their battles and follow the “survival of the fittest” path.

Second, AI itself can solve the innovation problem by applying AI to determine the AI innovation paths.

Third, it is important to provide information to industry to select and advance technology paths. It does not make sense to wait and allow the technology to find its path into adoption.

Fourth, the investment must be evaluated to ensure that it is indeed going into those technologies where the return on investment can be maximized. This handholding is important when things become tough.

2. THE FALLACY OF AVOIDANCE

The report highlighted a sensational finding that America was not ready for AI and then linked the conclusion to open up the bank so we can be ready.

Our position is different. We believe that the risk of not fixing the underlying issues and root causes that led to the previous warnings to be ignored is far greater than the risk of any foreign country overtaking America as a global leader.

But the report was not about fixing the underlying problems or root causes. It was about making a laundry list of things that needed to be improved and then suggesting some consulting-style fixes. America is faltering in AI and that government needed to allocate more funds to put the country back on track is the narrative that is certainly more accurate and reasonable than the OSTP's position, but we believe that the recommendations will not be helpful without an associated plan to fix the gaps. The gaps are in developing a solid and comprehensive plan and to stop viewing OSTP's plan as a national strategy. The OSTP must keep its focus on R&D and funding for federal AI. A new effort is needed to develop a national strategy, and it must be done outside of the OSTP.

3. CHINA IS IMPORTANT, BUT NOT REALLY

The initiative was clearly driven by a response to Chinese progress in AI and not so much to the organic development potential of AI in America. This lens of reactivity turned the final report into a document that seemed like a defensive strategy outlined to keep an invading army outside the city walls. Reminiscent of a plan to counter a siege, it lacked any forward thinking beyond fortifying the walls and fixing what's broken. There was no inspiration—just the basic defensive maneuvers. The China obsession was evident from the fact that the report uses the word “China” a whopping 604 times and “United States” 714 times. The word “America” (or “Americans”) appears 184 times while “Chinese” appears 91 times. The report, it seemed, was just as much about China as it was about America. Whatever the report was expected to accomplish, from the outside it seemed like the commission members were pulling the ropes hung from the belltower bells to warn the city that it is under siege. There was no excitement. There was no vision. There was no inspiration. Just a boring battle plan to ward off the attacking army and fortifying the fortress.

We believe that while that level of realism and pragmatic thinking is important, the country also needs inspirational messaging to motivate Americans to develop and embrace AI. This requires building an American AI-centric identity.

The one thing we are trying to point out in our book is that America does not need to be either in a passive mode or a reactive mode. America needs proper leadership that is neither based on fear nor on paranoia. What is needed more than anything is the inspirational leadership that can inspire and mobilize the nation. Instead, plan after plan and report after report, we are simply creating plans that people will read but that will not inspire grassroots movement or mobilize the nation.

4. BELIEVABILITY

The authors of the report did not consider the change management and national identity–shaping transformation that will be needed for America. It requires a delicate balance. Americans believe in their country and its potential. They believe in American exceptionalism. They have been told about America's magnificent performance, its leadership, its strength, its power, and its ability to overcome difficulties. To suddenly find out that the country may not be a leader in less than a decade or that the nation is experiencing strategic vulnerability may confuse people or create unnecessary hostility against minorities or infighting. It can truly affect the psychology of the nation. This can paralyze the spirit of the nation to come together to emerge as an AI nation.

Even if true, the narrative that all of a sudden a new competitor rose from the East that will use a technology to threaten our economy, society, political stability, spiritual well-being, ideologies, intellectual progress, and other pillars of our country may also score low in terms of believability for the American populace, and hence America may not mobilize to rise to the challenge. Even the terminology used in the report, “infiltrate in in our society, steal our data, and interfere in our democracy,” may sound extremely alarming for some people and not too threatening for others.

While every American was touched by the Internet, AI remains as an unknown. In the past American leaders have successfully inspired and motivated the nation. Current American leaders would need to discover a positive but pragmatic messaging to inspire the nation.

AI is not an easy technology to understand. Now consider that making the argument that AI is a powerful technology that will somehow tip the balance in China's favor (and by the way we have the world's most powerful tech companies) may not be believable for some people while it may really disturb other people, who may overreact.

So from a position of lacking trust, we push down on people a fairly complicated argument and expect people to be inspired by it. That's a recipe for disaster.

5. BREAKING DOWN THE SILOS

There are two major AI clubs in America. One could be described as the Silicon Valley club. The second as the DC club. Both have some similarities and some major differences.

Laying out the similarities between the two clubs is far easier than identifying the differences. Energized by the spirit of capitalism, both clubs are designed to create massive wealth for their members. Both clubs are trying to capitalize by developing AI capabilities for commercial and government clients. And both clubs are not necessarily approaching AI in terms of the national interests of the United States.

In terms of differences, the Silicon Valley club is composed of typical investors, celebrity-worshiping fans, plenty of foreign money, mostly West Coast–based academics and researchers, recycled management teams, dedicated nonobjective biased celebrity-pleasing media, tech celebrities and their fan clubs, and a unique culture.

The DC club is composed of investors (albeit of a different kind), academics typically from the East Coast universities, business leaders, retired military and government officials, former senators, governors, and congressmembers.

The Silicon Valley club's entrepreneurial constituency generally builds companies to fill the functionality and features gaps in the existing products of large firms. Occasionally they try to develop their own innovative business models, but as soon as their business models begin congealing, they are acquired by large tech firms. The goal of this club is simple: to create designer firms that can be positioned for sale to Big Tech.

The DC club is composed of systems integrators and government contractors who are also trying to embrace the new AI paradigm. To be accurate, due to the use of artificial intelligence in national security and weapons, the DC club was proficient in AI long before the field became prominent in Silicon Valley. The singular goal for this club is to win larger government contracts.

Caught in the midst of these two obsessions is the United States of America. These two clubs respected each other's space and kept a decent distance with each other, which allowed them to coexist. But the Silicon Valley firms started reading some obvious signs. As the economy was experiencing consolidation across major sectors, the growth opportunities were getting limited. Big Tech firms had two growth opportunities: first, to enter new sectors and redefine the competitive dynamics of those sectors with technology; and second, to go after the government business. Financial markets reward Big Tech on revenue growth, such that margins and revenue mix are often ignored. It did not matter where the revenue came from, what mattered was that they could show growth. And nothing beats the nice long-term government contracts.

Northern Virginia, Maryland, and areas in the DC suburbs suddenly saw a surge in Silicon Valley firms moving to the area.

The DC-area firms had developed powerful skills and domain knowledge to bring in new government business via RFPs and know how to navigate the complex world of government acquisitions (procurement). The government-focused technology sector is composed of both large firms (such as Boeing, Lockheed, GDIT, IBM, and Booz Allen), some midsize firms, and thousands of small firms. Many newly emerging AI firms—such as NCI—are now coming from the DC area. The nature of government contracting is such that it has created an ecological distribution parity where everyone gets something as long as they play the game. In this civilized jungle, the ecology calls for cooperation. The primes for major contracts could be small, so disadvantaged firms and larger firms would have to follow the lead of the small business. In other cases, large contracts awarded to giants necessarily require that some parts of the contract trickle to smaller firms. This natural order invites cooperation and has created a culture of cooperation. A far cry from the Silicon Valley's dog-eat-dog world where competitors are swallowed live and the game is played to eliminate each other, the DC club is designed to sustain each other. The DC ecology was about to get a culture shock as Silicon Valley firms began entering in the government space. AI is going to change the DC dynamics. Some of the DC-area firms have had long experience in AI. Their work on the national security and defense side required extensive machine learning work. Despite their performance record, they are not built like Silicon Valley firms.

On many AI RFPs the firms from the two clubs have to work together. Each brings some level of capability to win business. Both can learn from each other. This is something that should not be forgotten in the AI strategy. We believe change management of this type should be part of the national AI strategy.

Similarly, non-tech firms should not be excluded from the AI world. All AI planning must help move all sectors forward. This can be accomplished via engagement with scientific, professional, and trade societies.

Along the same lines, the acquisition and procurement staff of GSA and other agencies need to be retrained on how to procure AI. In December of 2021 the Senate passed a bill that directs agencies to retrain acquisitions staff. This is a good start. Congress should get far more involved in AI national strategy. The story of Congress is presented in the next chapter.

REFERENCE

1. NSCAI. 2021. “Final report: National Security Commission on Artificial Intelligence. NSCAP.” [Online]. Available at: https://www.nscai.gov/2021-final-report/.