On May 17, 2022, the Scowcroft Center’s Forward Defense practice collaborated with Applied Intuition, a Bay-area based tech startup, to host Nexus 22, a symposium at the intersection of defense and autonomy.
The event consisted of twelve panels, spotlight remarks, and keynote addresses by senior leaders including Michèle Flournoy, Marc Andreessen, Representative Mike Waltz, Representative, James Langevin, Chris Lynch, and Mike Brown, who came together to discuss issues at the intersection of autonomy, defense, and national security. The symposium addressed and answered the following key questions on the future of autonomous systems:
Why does the United States need to invest in autonomous systems?
Preparing for the conflict of the future. Despite speculation that the wars of the future would be fought differently, the Russia’s invasion of Ukraine has shown that some conflicts will rely on the same tanks and ground maneuvers that were prevalent in the twentieth century. Nevertheless, as Atlantic Council NRSF Margarita Konaev pointed out, we often conflate capabilities of the future with the technology that is “battle ready.” As Michèle Flournoy noted later in the day, the next 5-7 years will be a critical time for innovating and integrating these technologies, which many panelists suggested will be fielded by our adversaries during that time. Adversaries have been clear on their intentions to achieve next-generation capabilities first and dominate the defense space. The United States must keep pace.
What are the upshots to autonomous systems?
Sensing and deciding. Humans are not excellent decision-makers, notes Marc Andreessen. Taking an example from commercial autonomy, he noted that humans are distracted drivers. They text, drink, put on makeup, and multi-task. On the flipside, autonomous vehicles have 360-degree awareness at all times; they do not get drunk or put on makeup, and can multi-task, much more effectively than humans. Human in the loop, or HITL, autonomous systems, are a good starting point—and in some cases, may be the ethical choice, when it comes to lethal decision-making. But in many cases, the ethical choice may be autonomous systems that don’t get distracted.
What are the main pain points in fielding autonomous systems?
Speed. Technology is moving quickly—and US defense is falling behind. While the war in Ukraine and growing tension over Taiwan will help drive faster action, the Department of Defense (DoD) needs to also learn how to think and act more like private industry. After all, “five years in Silicon Valley is fifty years compared to everywhere else,” says Katherine Boyle of Andreessen Horowitz. By integrating commercial stakeholders and innovating with greater speed, the DoD can reduce procurement-to-capability timelines and compete with authoritarian systems such as China, where public-private partnership is woven (or coerced) into the institutional tapestry. Capitalism may not be a perfect system, but it can encourage technological breakthroughs. While the DoD should move with what Direcotr of Emerging Capabilities Policy Dr. Michael Horowitz calls “responsible speed,” the Department can still learn from these Silicon Valley’s mantra: Move fast. Break things
Communication. A dialogue between the public and private sectors must take place. On the public side, decisionmakers and warfighters will need to up their digital fluency, and understand emerging technologies, in order to work with them efficiently. While many private companies have shied away from DoD collaboration, panelists throughout the day noted that Silicon Valley engineers are increasingly patriotic and willing to work on dual-use capabilities. To work with the private sector, the DoD can help by educating private companies on growing threats and clearly communicate their intentions, which will build better understanding and foster lasting a partnership. Parts of private industry is caught up in a version of reality that doesn’t exist, rather than an increasingly multipolar world where the US is losing its competitive advantage. Communication—of intentions, of threats—will allow the public and private sectors to collaborate more closely, field autonomous systems faster, and deter future threats.
How can DoD accelerate cutting-edge capabilities?
Focusing in on the “meat” of acquisition. If US defense innovation is a sandwich, then it is missing the middle. The bread (and butter) of US defense acquisition is cutting-edge tactical concepts, and joint warfighting capabilities (think JADC2). However, cautioned Michèle Flournoy, the “meat” of campaign-level concepts is missing: how forces operate in any given theater, how tasks get assigned, how US forces amass resilience and redundancy—this critical middle component will be key to unlocking the full potential of emerging technologies and applying them to US defensive capabilities.
Managing talent. An essential component of bolstering US defensive capabilities is retaining and managing talent. Right now, the DoD does not effectively leverage STEM talent in order to make use of the STEM grads currently in service, nor does it effectively integrate outside talent into the DoD—which could look like, for example, creating a 100% digital reserve unit. Furthermore, the DoD does not effectively take advantage of private sector innovation, despite the fact that part of the entrepreneur ecosystem is driven by what Bilal Zuberi of Lux Capital called “the desire to protect the freedoms that we all cherish.” Improving on these pain points and building bridges with the private sector will allow DoD to develop existing talent and incorporate outside talent much more efficiently.
Measuring the right performance indicators. The number of tanks, planes, and ships, is the wrong qualitative paradigm for measuring US capabilities. Instead, measuring speed, accuracy, and resilience of decision-making and command and control (C2) can help benchmark DoD success and shortcomings. Artificial intelligence, machine learning, and human-machine teaming will help activate US capabilities and improve US warfighting.
How do we build trust in autonomous systems?
Testing & evaluation. Trust—or “justified confidence,” as it is known within DoD—is not so much built, as it is calibrated, says Jane Pinelis. Core to calibrating trust, is initial (and continual) testing and evaluation (T&E) designed to identify risks and improve success. Here, simulation can also play a big role, allowing for diverse use case testing before live trials.
Training and education of human operators, in tandem with academia and private industry, can also help establish confidence and reliance. Optimizing the talent acquisition and retention pipeline will allow DoD to up-level servicemember knowledge and advance autonomous applications.
Practical application. Finally, the practical application of autonomous and machine learning systems will help demonstrate emerging capabilities and build confidence in new technologies. Though it is important to mitigate unwarranted trust and operator complacency, repeated observation of autonomous systems in use will organically build trust at the public and operational levels.
Emerging technologies are more important than ever in the safeguarding of US values domestically and abroad. Building a bedrock of trust from the outset will help lay the foundations for asymmetric advantage in the long-run.
Defense keynote with Michèle Flournoy
What are the takeaways from the Russia-Ukraine war?
Russia is not ten feet tall. Systemic issues in authoritarian governments can affect how they wage war. Limits on dissent kneecap strategic advice; corruption siphons off resources; deficits in training and personnel inhibit combined arms campaigns and domain dominance. In the end, Russia’s “asymmetry of interest and resolve” contrasted sharply against the strong Ukrainian will to fight—and taken in totality, the Kremlin offensive so far has failed to achieve substantive gains.
Agility and tech matter. After the invasion of Ukraine, several NATO allies (including the US, Canada, and the UK) deployed special operators to help train the Ukrainian military in asymmetrical tactics, to use against a larger force such as Russia. During the war, Ukrainian troops have shown remarkable aptitude in nimbly integrating old technologies with new ones.
Sanctions must be strategic. Sanctions are a useful tool for effecting change without resorting to military force (read more on this “gray zone” of conflict here)—but they have their downsides. Though the war in Ukraine has pushed allies together, catalyzing well-coordinated action on sanctions, it has also pushed China to re-evaluate its dependence on the dollar. In the future, the US must calibrate sanctions to meet its goals without producing unpleasant side effects.
How has the Russia-Ukraine war mobilized the transatlantic community?
Sounding the alarm. The combination of Russian aggression in Ukraine with Chinese assertive behavior in the Indo-Pacific has awakened the transatlantic community. In particular, partnerships built around technology development will be a key area for transatlantic cooperation. Flournoy imagines “tech clusters” built around core technologies—and even a “T10” or “T12” (Similar to the G7, or Group of Seven, forum) framework to help foster collaboration and unity.
How can the US manage security in the Indo-Pacific?
Don’t fall asleep at the wheel. The US can’t rest on its laurels. Already, China is developing asymmetric capabilities that will challenge the traditional way of war. The US must start thinking more like the underdog, and focus on developing capabilities that will impose costs on China early on in a conflict where it will not have home-court advantage.
Prioritize the next 5-7 years. According to Flournoy, the next five-to-seven years will be critical. While Chinese president Xi Jinping is distracted by the COVID crisis, a struggling economy, and 2022 Party Congress, the US must act with speed to develop new capabilities to deter and defend against a new type of adversary. In the medium-term, as Xi begins to consider his legacy, he may also see a window of opportunity to incorporate Taiwan, while the US is still wondering where it put its wallet.
Venture capital keynote with Marc Andreessen
What is American Dynamism?
Building things. Buildings, bridges, dams; US cities; the United States itself—are all startups. At its core, “American Dynamism” is building big things. But since the 1960s, the United States has begun to move away from enterprise projects: it has become what Andreessen calls a “vetocracy,” or a system where a lot of people get a veto, and get to block new projects (such as housing units in San Francisco, or even hospitals). This means that less things are getting built, and the United States is becoming its own bottleneck.
Innovating. But the United States is also inherently resilient: a naturally defensible territory with a natural wealth of resources. Unlike China, the United States has the physical resources to become energy independent. The heavy lift lies not in tangible or intangible assets, but in breaking through rigid market structures and enabling innovation.
How has the Russia-Ukraine war shifted Silicon Valley bubble culture?
By bringing national security threats into focus. Silicon Valley has deep roots in defense and intelligence—but twenty years ago, a cultural rift began to form as the public-private technology flow inverted, and new technologies are deployed first to consumers, and then to the government. In many cases, that meant partnering with Chinese companies, who are obligated to share data with the Chinese military. The defense-commercial divide grew as Silicon Valley workers grew comfortable on the peninsula, and oblivious to growing global threats. The war in Ukraine is starting to move the needle again in the other direction as Americans become more aware of nascent threats to national security, and setting the playing field for small companies a chance to break back into defense. In addition to a Silicon Valley culture shift, the government is also shifting into higher gear and leaning in on emerging technologies, in order to prepare for twenty-first century threats.
How can small tech firms play to their strengths?
Innovating at pace. The US market, comments Andreessen, is not a free market system: it is an oligopoly entwined with government interests. Banking conglomerates, once deemed “too big to fail,” have grown even larger than they were at the time of the 2008 crash; new bank charters are at an all-time low; and a revolving door policy ensures that regulators and banking conglomerates are inextricably linked. But this presents an opportunity for agile startup businesses, who can innovate faster and better than bloated, restricted incumbents. If they can maneuver past big interests and hone in on unique opportunities, small companies can remain competitive in a choked market.
What is post-COVID market volatility doing to venture capital and tech startup funding cycles?
Putting them in a holding pattern. Right now, the market is in a holding pattern. Investors are still reeling from COVID-19 whiplash, where markets dropped, then rose, then dropped again—similar, noted Andreessen, to Wiley Coyote stepping over the edge of a cliff, hovering for an instant in mid-air. The volatility increases uncertainty: is this a temporary drop—in other words, a time to buy? Or is it a sustained downturn? The next six-nine months will help clarify larger trends, but, in the meantime, startups are raising capital at old prices, or not at all.