The 2018 National Defense Strategy outlines a comprehensive strategy comprised of multiple priorities, including readiness, modernization (nuclear, space and cyber, intelligence, surveillance and reconnaissance, missile defense, autonomous systems, and contested logistics), agility, and cultivating workforce tech talent.
AI’s Role in Defense Preparedness: Aircraft Readiness
The U.S. Air Force has about 5,600 aircraft with an average age of 28 years, some introduced into service 60 years ago. According to Air Force data, about 71.3 percent of the Air Force’s aircraft were flyable, or mission-capable, at any given time in fiscal 2017,23 which represents a drop from the 72.1 percent mission-capable rate in fiscal 2016, and a continuation of the decline in recent years.
In September 2018, the Secretary of Defense ordered the Air Force and Navy to raise mission-capable rates for four key tactical aircraft above 80 percent within a year, a number well above the mission-capability rates those aircraft now achieve:
• F-35 (readiness of 55 percent)
• F-22 (readiness of 49 percent)
• F-16 (readiness of 70 percent)
• F-18 (readiness of 53 percent for F/A-18E/F Super Hornet fleet, 44 percent for reserve fleet of F-18C Hornets)
Many of these aircraft are nearing the end of their service life. The Congressional Budget Office projects that replacing the aircraft in the current fleet would cost an average of $15 billion a year in the 2020s. That figure would rise to $23 billion in the 2030s and then fall back to $15 billion in the 2040s. In comparison, appropriations for procuring new aircraft averaged about $12 billion per year between 1980 and 2017.
The flying branch needs to replace a lot of planes in a short span of time, U.S. Air Force Secretary Heather Wilson said in a March 2018 statement: “The Air Force must manage a bow wave in modernization over the next ten years.”24
A significant improvement in aircraft readiness, along with the resulting cost savings (such as reduced maintenance and inventory holding costs), could have a major impact in bridging the modernization budget gap.
Improving aircraft readiness requires anticipating unplanned maintenance, ensuring availability of spare parts in the right locations, and scheduling maintainers to perform maintenance. The U.S. Air Force has demonstrated that by applying sophisticated AI algorithms to aircraft flight, environmental, and maintenance data, we can correctly identify at least 40 percent of unscheduled aircraft maintenance.
Based on the success of the demonstration project, the USAF plans to make the predictive maintenance application available to any USAF aircraft platform. When widely deployed, the USAF expects it will improve overall readiness by 40 percent.
This AI-based approach does not require the time and expense of retrofitting older jets with system sensors. These algorithms can be deployed in conjunction with the Air Force’s current maintenance systems to recommend additional aircraft subsystem maintenance while an aircraft is undergoing scheduled maintenance. The outputs of these algorithms can also inform spare parts demand planning and maintenance scheduling.
AI algorithms can not only increase the throw-weight (i.e., weapon delivery capacity) of the Air Force by 40 percent but also free up operations and maintenance budget for fleet modernization and expansion. The cumulative impact of this AI-driven predictive maintenance application for the USAF—which includes sizeable reductions in time, labor, spare parts and inventory holding costs, and reporting costs—adds up to a 10 percent efficiency improvement on $50 billion in annual operations and maintenance costs. That is $5 billion a year that can help bridge the modernization budget gap.
FIGURE 8.2
The Department of Defense’s Operations and Maintenance budget request for fiscal 2019 is $283.5 billion.25 Application of these AI algorithms across assets employed in the Army, Navy, and Marines can drive similar readiness improvements and budget savings.
Defense Modernization at the Speed of Relevance
U.S. Defense Department investments in AI provide the opportunity to upgrade military operational capabilities while improving affordability and streamlining key business functions to deliver military competitive advantage. Doing so at an accelerated pace and in partnership with innovative companies is a Defense Department priority.
The Department has made important strides in addressing the mismatch between the pace of start-ups and traditional government procurement through the creation in 2015 of the Defense Innovation Unit (DIU), a Silicon Valley–based office that has more flexibility to fund small defense contracts quickly and facilitate interaction between the Pentagon and Silicon Valley.
“My fundamental view is we are in a technology race. We didn’t ask to be in this, but we’re in it,” said Michael Brown, Director of DIU. “I’m concerned that if we don’t recognize that we’re in a race and take appropriate action, then we let China move forward and we don’t put our best foot forward in terms of leading in these key technology areas.”26
The Pentagon has also created a new Joint Artificial Intelligence Center to coordinate and advance defense-related AI activities. These efforts make it easier for the government to work with start-ups and other companies not used to working with the federal government.
In 2018 the Army established the Army Futures Command to streamline and accelerate acquisition and rapidly deliver warfighting capabilities, with an objective of reducing the requirements development process from nearly five years to one year or less.
With cross-functional teams, the command is tackling six modernization priorities: Long Range Precision Fires (missiles); Next Generation Combat Vehicle (tanks, troop carriers); Future Vertical Lift (helicopters); Network Command, Control, Communication, and Intelligence; Air and Missile Defense; and Soldier Lethality (troop combat capabilities).27
The fiscal year 2019 National Defense Authorization Act that was signed by the President in August 2018 earmarked $10.2 billion to help fund those efforts. The command itself has an annual budget of about $100 million.
“Army Futures Command has been an opportunity for us to bring two key stakeholders together—acquisitions and requirements,” said Under Secretary of the Army Ryan D. McCarthy during a panel at the 2018 Association of the U.S. Army (AUSA) annual meeting.28
The Defense Department is investing broadly in military application of autonomy, artificial intelligence, and machine learning. Efforts to scale Department-wide use of AI to expand military advantage include:29
Space and cyberspace as warfighting domains. The Department will prioritize investments in resilience, reconstitution, and operations to ensure U.S. space capabilities. The U.S. will also invest in cyber defense, resilience, and the continued integration of cyber capabilities into the full spectrum of military operations.
Resilient and agile logistics. Investments in this area prioritize distributed logistics and maintenance to ensure logistics sustainment while under persistent multidomain (ground, air, sea, and space) attack, as well as transitioning from large, centralized, unhardened infrastructure to smaller, dispersed, resilient, adaptive bases.
In the private sector, AI algorithms are in use to achieve similar goals such as high levels of on-time customer delivery and power network resiliency. The algorithms operate by continuously characterizing and compensating for uncertainty in demand, logistics, and supply networks; these techniques can be applied to ensure guaranteed delivery of military forces and supplies. Similarly, AI algorithms are used to track the state of power networks and ensure their resiliency; in the case of unanticipated equipment failure, power is instantaneously re-routed to ensure continuous power delivery. These techniques are equally applicable to operating in contested military environments.
Command, control, communications, computers, intelligence, surveillance, and reconnaissance (C4ISR). Investments prioritize developing resilient, survivable, federated networks and information systems from the tactical level up to strategic planning. Investments also prioritize capabilities to gain and exploit i
nformation, deny adversaries those same advantages, and enable the U.S. to identify and hold accountable both state and non-state perpetrators of attempted cyberattacks.
An example is the F-35 Joint Strike Fighter operational threat library—the “mission data files.” The mission data files—the “brains” of the airplane—are extensive on-board data systems that compile information on geography, air space, and potential threats such as enemy fighter jets in areas where the aircraft might be deployed in combat.30
The mission data files work with the F-35’s Radar Warning Receiver that detects approaching threats and hostile fire. Information from the aircraft’s long-range sensors is compared in real time against the library of enemy threats. If this happens quickly enough and at a sufficient standoff range, the F-35 can identify and destroy enemy targets before it is vulnerable to enemy fire. For example, the mission data system could rapidly identify a Chinese J-10 fighter if detected by the F-35’s sensors.
Keeping the mission data files current requires assimilating information gathered from numerous government intelligence agencies in varying formats (image, video, comments, documents, and structured data). Analysts must review, analyze, organize, and update these data manually and determine which intelligence is most relevant for a given mission data file. AI algorithms can be used to automate the aggregation, analysis, and correlation of these disparate data, which will accelerate the process and ensure the F-35 is operating on relevant and up-to-date information.
Workforce AI Talent
Increasing the pipeline of AI talent in both the private and public sectors is critical to ensuring AI leadership and national security. AI researchers and data scientists currently command increasingly high salaries, a sign of the shortfall in top-tier talent. The U.S. must educate, recruit, and retain the best researchers in the world. Increasing education funding and opportunities for science, technology, engineering, and mathematics (STEM) is a top priority. The federal government plan for STEM education released in December 2018 takes some important steps to expand the pipeline of students acquiring advanced STEM degrees.31
Another significant obstacle to progress is the security clearance process. There are 500,000 new federal employees, contractors, and military personnel whose ability to work is impeded or limited by a security clearance process that is unable to keep up with demand.32 Some individuals are able to obtain interim security clearances and begin working, but others are left waiting a year or more before being able to begin work at all.33 There is a similar backlog of reinvestigations that are required periodically of cleared personnel.
Defense Security Services is implementing innovative techniques to assess personnel risk, including a “continuous evaluation” (CE) system. Using AI algorithms, the CE system regularly scans a range of data sources—such as court proceedings and bank and credit bureau records—looking for indications that an existing clearance holder or applicant may pose a security risk. If it uncovers any risk indicators for an individual, the CE system alerts a security agent for follow-up review and action.34 These systems are more efficient and more effective at identification of risk and can easily be scaled to federal employees and suppliers.
No Bigger Stakes
AI will fundamentally determine the fate of the planet. This is a category of technology unlike any that preceded it, uniquely able to harness vast amounts of data unfathomable to the human mind to drive precise, real-time decision-making for virtually any task—including the operation of hypersonic missiles; energy-powered space-based weapons; autonomous air, land, and water combat vehicles; and advanced cyber-weapons aimed at information systems, communications networks, power grids, and other essential infrastructure. In short, AI will be the critical enabling technology of 21st-century warfare.
Today the U.S. and China are engaged in a war for AI leadership. The outcome of that contest remains uncertain. China clearly is committed to an ambitious and explicitly stated national strategy to become the global AI leader. Unless the U.S. significantly steps up investment in AI across the board—in government, industry, and education—it is at risk of falling behind. You can think of this as the 21st-century equivalent of the Manhattan Project.
The fate of the world hangs in the balance.
Chapter 9
The Digital Enterprise
As we’ve covered in the preceding chapters, the confluence of elastic cloud computing, big data, AI, and IoT drives digital transformation. Companies that harness these technologies and transform into vibrant, dynamic digital enterprises will thrive. Those that do not will become irrelevant and cease to exist. If the reality sounds harsh, that’s because it is.
The price of missing a transformational strategic shift is steep. The corporate graveyard is littered with once-great companies that failed to change. Blockbuster, Yahoo!, and Borders stand out as companies crushed by sector-wide changes to which they could not adapt. At its peak, Blockbuster, an American video and videogame rental company, employed 60,000 people, earned $5.9 billion in revenue, and boasted a $5 billion market capitalization.1 Six years later, Blockbuster filed for bankruptcy, a shell of its former self.2 In 2000, Netflix CEO Reed Hastings proposed a partnership with Blockbuster—Netflix wanted to run Blockbuster’s online presence as part of a $50 million acquisition—and Blockbuster declined.3 As I’m writing this, Netflix has a market capitalization of over $160 billion and Blockbuster no longer exists. Netflix saw the shift happening, discarded mail order, and transformed into a streaming video company. Blockbuster did not.
In 2000, Yahoo! was the poster child of the internet. It was valued at $125 billion at the height of the dot-com bubble.4 In the following years, Yahoo! had opportunities to buy both Google and Facebook, and in both cases, they failed to execute the deals due to pricing issues: $3 billion was simply too much to pay for a company like Google.5 In 2008, Microsoft attempted a hostile takeover of Yahoo! at a price tag of about $45 billion, which Yahoo! successfully rebuffed, and in 2016, Verizon acquired Yahoo! for $4.8 billion.6 The downhill slope for Yahoo! was dramatic. The consumer internet became mobile, social, and driven by interactions around photos and videos. Yahoo! did not, so it was acquired and disintegrated. At the time of this writing, Google and Facebook have market capitalizations of $840 billion and $500 billion, respectively.
Borders, an American book retailer, had 1,249 stores at its peak in 2003.7 Just two years earlier, Borders contracted its e-commerce business to Amazon.com, a famously grave mistake that disincentivized internal executives from establishing a proprietary online presence.8 Amazon’s e-books and data-driven logistics were an existential threat, but Borders failed to act. Unsurprisingly, by sending its online business to Amazon and failing to enter the e-book business, the Borders brand eroded and it became irrelevant. In 2010, Borders attempted to launch its own e-reader and e-book store, but it was too late.9 One year later, Borders closed its doors for the last time.10
These stories—Blockbuster, Yahoo!, and Borders—are not exceptional. They are not anomalies. They are not unusual. These stories are the product of mass corporate extinctions driven by fundamental shifts in the way business is done. They represent the guaranteed outcome for companies that do not transform. This time around, digital transformation is the do-or-die impetus. Companies that do not succeed in this vital task will go the way of Blockbuster, Yahoo!, and Borders.
While the cost of failing to adapt is perilous, the future has never looked brighter for large companies embracing digital transformation. This is primarily for two reasons. The first is Metcalfe’s Law, as we have discussed in chapter 7: Networks grow in value as the participants increase. Large corporations stand to benefit from a similar paradigm regarding data. If properly used, the value of enterprise data also increases exponentially with scale. Large companies tend to have dramatically more data than the upstart competitors seeking to supplant them and can collect data considerably faster. If incumbent organizations can digitally transform, they will establish “data moats,
” which are an asymmetric advantage that could dissuade competitors from easily entering their industries. The impact of such data moats should not be underestimated. The advantage Amazon or Google already has, thanks to years of consumer and user data, is enormous. Similarly, early market entrants with disruptive offerings—think Uber, Zappos, Slack, or Instagram—quickly gain competitive advantage by capturing and using large amounts of data, creating a formidable time-to-market scale advantage.
The second reason large companies are well positioned to exploit digital transformation is they typically have access to substantial capital. Digital transformation offers highly attractive investment opportunities. One such opportunity is hiring large numbers of top-flight data scientists and engineers. Another is investing in digital transformation technologies.
As it turns out, these two factors—data moats and access to capital—work synergistically. Large companies with proprietary data, the right technologies in place, and the capital to recruit top talent will find themselves in almost unprecedented—and extremely favorable—positions.
For data scientists and engineers, more data means more interesting problems to solve—and the best data scientists and engineers want to work on the most interesting problems. If a large company is able to execute a successful digital transformation, its data moat—i.e., its advantage in having more and better data than competitors—translates into the ability to attract superior data scientists, build superior artificial intelligence algorithms and outputs, generate superior insights, and ultimately achieve superior economic performance. Google, Amazon, and Netflix are clear examples of how such a data advantage plays out.
We are in the very early days of digital transformation. As we’ve seen, the technologies driving this change have matured only within the last 5 to 10 years. In this chapter, I share case studies of successful digital transformation initiatives at six large organizations—ENGIE, Enel, Caterpillar, John Deere, 3M, and the U.S. Air Force—based on work we do at C3.ai with these customers.
Digital Transformation Page 18