THE DAWN OF ARTIFICIAL INTELLIGENCE

On November 30, 2016, Senator Ted Cruz organized America's first Senate hearing on Artificial Intelligence. It was named “The Dawn of Artificial Intelligence” (Cruz 2016). The hearing invited significant interest. Lines started forming outside the hearing room, and when the gates opened, all seats were taken. In a jam-packed room, the hearing began with Senator Cruz introducing the AI revolution. In his introductory remarks, he mentioned, “Many believe that there may not be a single technology that will shape our world more in the next 50 years than artificial intelligence. In fact, some have observed that as powerful and transformative as the Internet has been, it may best be remembered as the predicate for artificial intelligence and machine learning.”

Sen. Cruz had invited five witnesses to the hearing, including Greg Brockman, cofounder and chief technology officer, OpenAI; Dr. Steve Chien, senior research scientist, Autonomous Space Systems Technical Group supervisor, Artificial Intelligence Group at NASA's Jet Propulsion Laboratory, California Institute of Technology; Dr. Andrew Futreal, professor in the Department of Genomic Medicine at the University of Texas MD Anderson Cancer Center; Dr. Eric Horvitz, interim co-chair, Partnership on Artificial Intelligence and managing director, Microsoft Research Lab; and Dr. Andrew Moore, dean of the School of Computer Science at Carnegie Mellon University.

Senator Cruz was up-front and clear about what was at stake. He stated:

Today, the United States is the preeminent leader in developing artificial intelligence. But that could soon change. According to the Wall Street Journal, ‘the biggest buzz in China's Internet industry isn't about besting global tech giants by better adapting existing business models for the Chinese market. Rather, it's about competing head-to-head with the U.S. and other tech powerhouses in the hottest area of technological innovation: artificial intelligence.’

Ceding leadership in developing artificial intelligence to China, Russia and other foreign governments will not only place the United States at a technological disadvantage but it could also have implications for national security.

We are living in the dawn of artificial intelligence, and it is incumbent that Congress and this subcommittee begin to learn about the vast implications of this emerging technology to ensure that the United States remains a global leader throughout the 21st Century. (Cruz 2016/Ted Cruz).

Senator Cruz's words showed that he was fully cognizant about the Chinese desire to compete head-to-head with America in AI and dethrone America as a leader. He warned about the results of the potential loss in the leadership position. In no uncertain words, he warned America that the nation must remain a global leader. This begs several questions:

• If it was known that AI was such a powerful advantage and mission-critical technology, then why did we not see a far more engaged Congress about this? Why were only one or two senators talking about it?
• Why was it not considered that the technology needed to be introduced to America in a positive and easy to understand manner, where masses could relate to it and get inspired by it?
• Notice that the witnesses came from the technical side. Why was there no representation from the policy or economic history side? Why were there no AI industrialization experts? Even though the focus of the investigation was on policy implications and impact on commerce, why were there no economists? As a side note, OpenAI, which was at that time a nonprofit, and was trying to position itself as an “AI for good” nonprofit, participated among others. Later, OpenAI became for profit and ditched the cloak of nonprofit.
• Why was the hearing so focused on the risks of AI without having a corresponding focus on the power of AI to transform the economy?
• Since most of the witnesses focused on three things—(1) increase funding, (2) technology, and (3) governance—why were other critical factors not considered as relevant?

Despite having such strategic clarity about AI as early as 2016, the American AI Initiative lost momentum in the later years. Five years later, it became clear that America had lost its position to China. Whatever Senator Cruz was trying to achieve did not work. The testimonies of the witnesses were unable to save America's leadership. One of the authors of this book attended the hearing and came out of it certain that without a corresponding execution plan, America would not be able to protect its competitive position.

The problem lay in listening to the technologists and not to innovation diffusion experts, industrialization experts, and technology historians. Technologists will simply state the obvious about how they see the world. In their world, the problem is about the technology they are working on, the problem they are trying to solve, about the data and the algorithms, and most importantly about the funding. In their world, if they can get enough money and resources, they can solve the technical problem. The question is, then what? Even if the technologists solve the technical problem, that does not mean that innovation will find its way in society or get implemented to maximize returns. All it means is that the researcher will be able to address the research problem.

You can call one technical expert or a million, the story would have been the same. More data sets, more money, more grants, more researchers, and more governance. The last one, governance, mostly for virtue signaling reasons to appear responsible—since responsibility was in fashion.

And the storyline during the hearings stayed the same. It was: AI is big, AI needs investment to maintain American leadership, give us more money, and it will build the American economy. The senators really did not require testimony from expert witnesses. The story line was standard and known in advance. Senator Cruz's own opening statement was probably far more insightful than any of the witness testimonies. No technologist could have added value at a national level without an associated innovation diffusion and industrialization plan. America did not need testimonies from technical experts and computer science department heads. America needed industrialization experts.

DIGITAL DECISION-MAKING

A year after Senator Cruz's hearing, on December 12, 2017, another US Senator, Roger Wicker (R-Miss.), who was chairman of the Subcommittee on Communications, Technology, Innovation, and the Internet, convened a hearing titled “Digital Decision-Making: The Building Blocks of Machine Learning and Artificial Intelligence.” He invited Dr. Cindy Bethel, associate professor, Department of Computer Science and Engineering, Mississippi State University; Mr. Daniel Castro, vice president, Information Technology and Innovation Foundation; Ms. Victoria Espinel, chief executive officer, The Software Alliance; Dr. Edward Felten, PhD, Robert E. Kahn Professor of Computer Science and Public Affairs, Princeton University; and Dr. Dario Gil, PhD, vice president, IBM Research AI and IBM Q, IBM. Once again, the entire panel was composed of technology experts. And as expected, the core theme of the witness testimonies turned out to be the same: AI is big, more funding is needed, and good things will happen with more money. Within all that, experts also talked about the advancement in deep learning and other innovations, but the main point always came down to asking for larger grants.

It is hard to imagine what additional insights were obtained by Senator Wicker—except precious time was being lost. That was the time when America should have been working full-steam ahead to capture the AI industrialization edge—as China was doing. That was when the technology adoption should have been worked on as part of a larger industrialization transformation. That did not happen.

China was working with a very different plan. Research was being approached from an industrialization perspective and industrial applications from research. A symbiotic relationship between research and industrialization was incorporated with a structured and organized approach. Investment was being allocated wisely and systematically. Unlike other technologies, machine learning applications develop from data. More applications lead to more data and more data to more applications.

Besides those hearings, other AI-related initiatives were attempted in Congress.

HOUSE GETS INVOLVED

In July of 2017 Congressman Delany published an article on The Hill after launching in May of 2017 a congressional caucus on artificial intelligence. Congressman Delany started his article by:

One of the biggest problems with Washington is that more often than not the policy conversation isn't grounded in the facts. We see this dysfunction clearly on technology policy, where Congress is largely uninformed on what the future of artificial intelligence (AI) technology will look like and what the actual consequences are likely to be. In this factual vacuum, we run the risk of ultimately adopting at best irrelevant or at worst extreme legislative responses. (Delaney 2017b)

Congressman Delaney, a former entrepreneur, represented a Maryland district. His words are extremely powerful as he claims that legislative and policy discussions are not based on facts, and that is especially true when it comes to technology. He proposed that Congress must take a proactive role and established a caucus. He received support from Rep. Pete Olson, a Republican from Texas, making the caucus a bipartisan undertaking. Several other members of Congress joined the caucus. At the final tally, a total of 28 members of Congress joined it. The representation was from across the country.

Besides congressman Delaney, the caucus included the following: Pete Olson (TX-22), Jerry McNerney (CA-09), Don Beyer (VA-08), GK Butterfield (NC-01), André Carson (IN-07), Emanuel Cleaver II (MO-05), Suzan DelBene (WA-01), Mark DeSaulnier (CA-11), Nanette Diaz Barragán (CA-44), Debbie Dingell (MI-12), Anna G. Eshoo (CA-18), Bill Foster (IL-11), Josh Gottheimer (NJ-05), Pramila Jayapal (WA-07), Henry C. “Hank” Johnson (GA-04), Ro Khanna (CA-17), Derek Kilmer (WA-06), Brenda Lawrence (MI-14), Ted Lieu (CA-33), Dan Lipinski (IL-3), Michael McCaul (TX-10), Bobby Rush (IL-01), Brad Sherman (CA-30), Darren Soto (FL-09), Elise Stefanik (NY-21), Steve Stivers (OH-15), and Marc Veasey (TX-33).

These congressmen and congresswomen included some of the most influential members in Congress. For example, congressman Don Beyer is chairman of the Joint Economic Committee, the Subcommittee on Space, and the Subcommittee on Research and Technology; and Dr. Bill Foster is a PhD in physics from Harvard. The sad part is that despite having a caucus composed of such a powerful group of people, the caucus did little to move AI forward. By the end of 2021, Congressman Delaney, Representative Olson, Representative Lipinski, and Representative Stivers were no longer in office.

The caucus was active between 2017 and 2018—and one year after coming together, the caucus was pretty much abandoned by the representatives. A website was created to report the activities of the group. A total of four press releases and six news/blog entries were made by the caucus. The last entry appeared in December 2018.

Representative Olson authored a blog entry and stated:

The AI Caucus is the perfect forum for legislators to be educated on the applications of AI in various applications, understand the issues and policy impacts surrounding AI and to ask the hard questions concerning the matter. The caucus provides an opportunity to learn the benefits as well as gauge the risk versus reward in this emerging field. (Olson 2017)

Both Rep. Olson and Rep. Delaney referred to a comment made by Elon Musk in which the famous entrepreneur warned the world that AI is a potential existential risk for humankind. Neither one of them questioned the validity of that comment or considered the fact that if Mr. Musk actually believed that, then why is he building a company that specializes in AI. Another possible explanation could be that Mr. Musk was in a creative mood when he stated that. He has been seen in different moods and even gave an interview while smoking marijuana. He was also fined by the SEC for making irresponsible statements. Both congressmen seem to have taken the warning seriously and focus of the caucus turned to governance, control, and safety.

Congressman Delaney then conducted a few interviews—in one of which he interviewed a Stanford professor who was also an entrepreneur. Congressional hearings took place in late 2017 and in February 2018. Congressman Delaney once again warned the nation about AI. This time his tone was different, and one could see that the overall anxiety was increasing. Referring to deficit spending, he said, “It's as predictable a crisis as we've ever seen. This is almost certainly going to happen.” And then he connected it with AI and said, “AI has the potential to make people healthier, create new jobs and connect communities. Government may not need to step in with regulations, but it will need to play a role in how these new innovations shape society. If we don't do it for this next wave of change, I fear the results will be very negative” (Delaney 2018).

Once again, we want the readers to understand the criticality of what was at stake. America was experiencing a powerful adversary. The primary mode and basis of competition was AI technology. The competitive loss implied significant competitive threat to the nation in terms of economic performance and in almost all other areas. More than any other time, that is when America needed a powerful and clear leadership from Congress—but that did not happen between 2016 and 2019, the critical and formative years of AI.

What is truly surprising is that the caucus seemed to have come to a sudden halt in December of 2018. The last press release was of Congressman Olson recognizing Congressman Delaney's contribution. Now that website stands as a relic in a museum reminding America that just when China was accelerating its AI adoption, expanding its footprint, developing AI with a high level of discipline, the American leadership was abandoning the caucus formed to focus on AI. The story of this abandonment will be preserved on the Internet. It became the story of withdrawal and relinquishment. A website will remind us of the good old days when there was hope. Now desolate, barren, and unattended, the site clings on as a reminiscence of a period when at least a bunch of congressmen and congresswomen attempted to make a difference. Both Olson and Delaney are no longer in Congress at the time of this book's publishing. Both left their mark—but what they did was not enough to move America forward in AI.

KILLER ROBOTS OBSESSION

AI is a complex technology. Unless you allocate some reasonable time to understand how machine learning works, it is counterintuitive to understand how a system rises out of data. But there is an easy part of AI. That is the governance, control, ethics, values, and killer robots part. While most people may not understand the functional aspects of AI, it is easy to view AI as some type of a terminator scenario. Hence, one way people can feel important and be part of the broader AI debate is by focusing on the killer robot scenario. By doing that you are in the limelight since you are talking about something that America already understands. And you can tap into the existing molds of monsters shaped over decades by Hollywood. If you are a congressmember, you can show your constituents that you care about them and the future of the world by being part of the “responsible” movement and saving them from the terminator robots. It was easy and it was popular. Hence, the overwhelming focus of Congress remained on the governance and ethics of AI.

Yet even as scandal after scandal erupted from Big Tech firms and their CEOs were pulled into hearings, no governance and ethics standards and frameworks were universally adopted by these firms. Despite so much attention being paid to governance, all the problems related to governance and ethics—for example, exploitation, manipulation, profiling—continue in Big Tech, but the government has not been able to influence them to change their ways. This shows that the entire governance and ethics posturing is superfluous and insincere.

So the governance and ethics posturing is problematic on two ends: first, it creates unnecessary and preemptive fear, and second, it does not solve the real problems of AI governance and ethics. It stays as virtue signaling and posturing.

The Senate Caucus

Nearly two years after Representative Delaney's failed attempt to mobilize the House Caucus, in March of 2019 Senator Heinrich and Senator Portman announced the formation of the AI Caucus for the Senate. Senator Heinrich said:

I am proud to announce the formation of the bipartisan AI Caucus in the Senate with Senator Rob Portman. Together, we will work with our colleagues to develop smart policy in a responsible way to ensure the United States remains at the forefront of innovation while maintaining important ethical, safety, and privacy standards. In the years ahead, it will be critical for policymakers to strike the right balance in developing the technology so that academia, private industry, federal agencies, and our national labs can harness the enormous potential of AI to the benefit of society and the American people. (Heinrich 2019b/Martin Heinrich)

This was an important milestone and much-needed change. One can ask several questions about this. First, when Senator Cruz conducted the hearing on AI in November 2016, he pretty much expressed a similar sentiment. Why did it take the Senate more than two years to launch this important caucus? Also, the Senate had the example of the AI caucus from the House. Again, why did it take that long? Those were the critical formative years of building AI. Second, why did we continue to belabor the same points over and over again: AI is important, we must harness its potential. The whole world knows that, and America needed to see some action.

After the formation of the caucus, it was as if a new level of consciousness and awareness had transpired at the Senate level. A month later, in a hearing before the Senate Committee on Armed Services, Senator Heinrich questioned then-US Secretary of the Air Force Heather Wilson and then-Chief of Staff of the Air Force General David L. Goldfein about how the Air Force, and the Department of Defense as a whole, were working to address the growing workforce needs as the military implements artificial intelligence initiatives. In the hearing, Senator Heinrich said, “Artificial Intelligence is going to be critical both on and off the battlefield. As you know, AI is not possible without good quality data, and AI is only effective if we have a workforce that understands how to take care of that data” (Heinrich 2019c). Here the senator was raising two important questions—first about the data science expertise and the second about the data. Secretary Wilson responded and acknowledged that there will be a huge shortage of data scientists and analysts over several decades and that it is a priority.

In response to the findings from the session, on May 16, 2019, Senators Heinrich and Portman introduced the bipartisan Armed Forces Digital Advantage Act, to modernize the Department of Defense (DoD) workforce by adding a recruitment focus and establishing military career tracks for digital engineering.

This showed what needed to get done in America. Ask the relevant questions, understand the problem, propose a solution, and then act on it quickly. Within two months America gained more velocity than in the last two years. The actions taken were commendable.

While the questions asked by Senator Heinrich and the actions taken were extremely relevant, the real answer lies not in hiring another HR professional to drive a wave of recruitment to hire more data scientists. Those are all tactical interventions, and they are important. However, to really address the question we must consider how to change the basic fabric of education in the nation so that the country produces data management and data science experts. The question is how the country can develop a national data strategy where it is understood what type of data will be captured, how, and why—and who will capture it, process it, and use it to develop new AI.

To address such concerns, in May of 2019, Senator Heinrich, Senator Portman, and Senator Brian Schatz (D-HI) proposed the National Strategy for Artificial Intelligence and called for $2.2 billion investment in education, research, and development. The three senators introduced S. 1558, the bipartisan Artificial Intelligence Initiative Act (AI-IA), which would organize a coordinated national strategy for developing AI and provide a $2.2 billion federal investment over five years to build an AI-ready workforce. The goal of the bill was to accelerate the responsible delivery of AI applications from government agencies, academia, and the private sector over the next 10 years.

In his introductory statement, Senator Heinrich said:

Artificial intelligence—and the opportunities and challenges it will bring—are becoming seemingly inevitable. Now is the time to formulate AI workforces and policies to keep these innovations on a responsible path. The Artificial Intelligence Initiative Act would ensure that the United States establishes a national strategy for AI research and development and would invest $2.2 billion in R&D. Whether it's Silicon Valley, New Mexico or Ohio, or elsewhere across the United States, our high-tech companies, universities, and national laboratories are the foundation by which we lead and maintain an advantage in AI. But the gap is closing quickly, and the United States will only continue its leadership position if it acts with a sense of urgency and purpose. If we defer AI development to other nations, important ethical, safety, and privacy principles will be at risk, which not only harms the United States, but the international community as a whole. When a new AI advancement is made—the AI-IA will have already put into place policy, people, and potential for this new market to truly take transportation, health care, manufacturing, and national security to the next stage of opportunity. (Heinrich 2019a/Martin Heinrich).

This was a good initiative. It recognized the need for urgency. In February of 2019, President Trump had already signed the American AI Initiative, and the OSTP had floated what they were calling the national strategy for AI. The above statement from Senator Heinrich seemed to indicate that a national strategy did not exist and that there was a need to establish one—including one for R&D. Was the government not talking to each other or Senator Heinrich did not consider the OSTP-led national strategy as a legitimate full-blown strategy for the country? We don't know the answer to this question, but we assume that the Senate was in action mode and had little patience for 2016–2019 indolence. The bill asked the president to establish and implement an initiative with respect to artificial intelligence to be known as the “National Artificial Intelligence Research Development Initiative.” This is something President Trump did in the last weeks in office.

According to the information about the bill from Senator Heinrich's office, the Artificial Intelligence Initiative Act accomplished the following:

• Established a National AI Coordination Office (director and staff to coordinate federal AI efforts), an AI Interagency Committee (senior leaders across federal departments), and an AI Advisory Committee (non-governmental experts) to develop a National Strategic Plan for AI R&D and to facilitate coordination across government agencies.
• Required the National Institute of Standards and Technology (NIST) to identify metrics that may be used to establish standards for evaluating AI algorithms and their effectiveness, as well as the quality of training data sets ($40 million each year).
• Required the National Science Foundation (NSF) to formulate educational goals for addressing algorithm accountability, explainability, data bias, privacy, and societal and ethical implications of AI. The NSF will also request to fund research on both the technical and educational aspects of AI and to analyze the AI effect on American society through awarding up to five new “Multidisciplinary Centers for Artificial Intelligence Research and Education.” At least one of these five centers will have K-12 education as its primary focus, one will be a minority-serving institution, and all will include a lifelong education component ($500+ million total; five centers at $20 million per year for 2020–2024).
• Required the Department of Energy (DOE) to create an AI research program, build state-of-the-art computing facilities that will be made available first and foremost to government and academic AI researchers but will also be available to private sector users on a cost-recovery basis as practicable. The bill established up to five Artificial Intelligence Research Centers to include institutions of higher education and national laboratories ($1.5+ billion; five AI Research Centers at $60 million each per year for 2020–2024).

In 2020, Senators Heinrich and Portman introduced another bipartisan AI-related bill. This time it was for the Armed Forces Act to strengthen the Department of Defense's AI capacity by increasing the number of AI and cyber professionals in the department. Perhaps the most important contribution from these senators came in the form of obtaining advancements for AI in the National Defense Authorization Act for Fiscal Year 2021 (FY21 NDAA). Some of the previous versions of the legislations championed by Senators Heinrich and Portman were rolled into this bill, including:

• Artificial Intelligence Initiative Act (AI-IA);
• Artificial Intelligence for the Armed Forces Act;
• National AI Research Resource Task Force Act; and
• Deepfakes Report Act.

Pulling In Other Bills

The FY21 NDAA included a modified version of the Artificial Intelligence Initiative Act (AI-IA) aimed at boosting US leadership in AI research and development. The AI-IA included in the NDAA was tasked with creating a National Artificial Intelligence Initiative Office to coordinate ongoing AI R&D and demonstrate activities among civilian agencies, DoD, and the intelligence community.

The bill also established the National Artificial Intelligence Advisory Committee. This committee will provide expert advice to policymakers and the National AI Initiative Office. The legislation also tasked the National Science Foundation with examining how the present and future US workforce can better prepare for and integrate AI systems.

The bill directed the National Institutes of Standards and Technology (NIST) to develop and advance collaborative frameworks, standards, guidelines, and associated methods and techniques for AI. Such development of technical standards and guidelines were expected to promote trustworthy AI systems. Overall, the bill authorized $1.5 billion over the next five years for these critical AI initiatives.

The conferenced NDAA included some provisions from S. 3965, the Artificial Intelligence for the Armed Forces Act of 2020, previously introduced by Heinrich and Portman. The bill was aimed at advancing the Department of Defense's AI capabilities on two fronts: (1) requiring new hires in armed services to take additional tests to determine computational skills, which was an indication that the future of the military was about automation; and (2) guiding the armed forces to hire and develop AI professionals and data science experts.

NDAA also included a version of the National AI Research Resource Task Force Act (S. 3890) previously introduced by Senators Heinrich and Portman. The provision in the FY21 NDAA were to organize a task force to develop a detailed roadmap for the development of a national cloud computer for AI research. It was expected that the cloud would now enable researchers and entrepreneurs around the country to access supercomputing capabilities, which were previously restricted to national laboratories and universities. This will result in building a national research cloud.

Lastly, the conferenced NDAA included a version of S. 2065 originally introduced by Heinrich and Portman. The provision in the FY21 NDAA required the Department of Homeland Security to assess and report on the state of digital content forgery technology—commonly known as “deepfakes.” This will help identify and assess the tactics employed by foreign governments, and their proxies and networks, for deepfakes to harm US national security.

In May of 2021, Senators Heinrich and Portman announced another bipartisan legislation. It was the Artificial Intelligence Capabilities and Transparency (AICT) Act. The AICT Act was proposed to implement recommendations of the National Security Commission on Artificial Intelligence's (NSCAI) final report. AICT's goal was to improve talent and accelerate adoption by agencies.

The legislation gave new authority and resources to the Department of Defense, Department of Energy, intelligence community, and Federal Bureau of Investigation in order to ensure the federal government is positioned to make the best use of rapidly evolving AI capabilities.

In May of 2021, Senators Heinrich and Portman also sent a letter to Director Sethuraman Panchanathan of the National Science Foundation, calling on him to ensure that the US remains a leader on AI, to establish AI institutes focused on ethics and safety, and to ensure that AI advancement is transparent, reliable, and in line with American values.

In October of 2021, Senator Heinrich appreciated the National Science Foundation (NSF) establishing a new theme prioritizing trustworthy AI as part of its National Artificial Intelligence (AI) Research Institutes. Senator Heinrich released the following statement:

If our country is to reap the benefits of AI at scale, we must ensure that citizens and users—human people—have the trust and confidence in AI systems to actually deploy them. That's why I'm pleased to see the National Science Foundation embrace the need to prioritize AI research on safety and ethics and increase efforts to establish a new AI research institute theme focused on studying and enhancing the trustworthiness of AI. In recent weeks, internal research of major tech companies like Facebook have revealed that AI systems can have unintended consequences on its users, including significant ethical and safety issues. Funding this research in the public domain will go a long way to help address those problems. I look forward to providing congressional oversight to ensure NSF puts sufficient resources towards this critical Trustworthy AI research topic. (Heinrich 2021/Martin Heinrich)

The efforts by the Senate were showing results. America was finally getting to the point where the country needed to be in 2016. Was it too little, too late? Only time will tell. The problem was that instead of leaving the OSTP-operated failed 2016–2019 leadership, the new AI Initiative outlined in the William M. (Mac) Thornberry National Defense Authorization Act for Fiscal Year 2021 (to accompany H.R. 6395) ended up back with the OSTP. If the past is any indication of the future, it is unlikely that the decline of American AI will be stopped, let alone reversed. The anchoring bias of the 2016 plan was now deeply ingrained and rooted in the nation. Senators were working hard with what they were given. The agencies were adapting and trying to make sense out of a delusional national strategy on one end and a flurry of legislation that were hitting them on the other end—all while trying to do their best to serve America during the Covid times.