In December, physicists at Hefei’s University of Science and Technology of China published a paper in Science with huge national-security implications. They claimed, as the journal Nature put it, “to have made the first definitive demonstration of ‘quantum advantage’—exploiting the counter-intuitive workings of quantum mechanics to perform computations that would be prohibitively slow on classical computers.” It was a major advance in the global competition to develop powerful quantum-computing capabilities that could, among other things, enable adversaries to break encryption algorithms and potentially undermine the security of all data sent over the internet, from government communications to financial transactions.
The threats go well beyond quantum computing. US adversaries including China, Iran, and Russia are investing in additional technological capabilities to counterbalance the United States’ advantages in great-power competition—specifically its dominance in kinetic operations and weapons. They are focusing on research and development (R&D) of dual-use technologies such as advanced computing, artificial intelligence (AI), and biotechnology, which will require the United States to focus on non-kinetic defense strategies to secure an advantage.
The 2018 National Defense Strategy (NDS) recognizes the impact that rapid technological advancements have on the security environment and acknowledges that the commercial sector develops some critical technologies. But it neglects to highlight the varying security advantages provided and risks posed by different technologies and instead lumps them all into the same strategy. The over-generalized approach puts the United States in an unfavorable position. It has contributed to challenges in engaging with the private-sector and civil-society actors, led to inadequate research funding for certain technologies, and more broadly resulted in a less cohesive strategy for researching and developing crucial technological capabilities.
The next NDS should include sophisticated, nuanced strategies for emerging technologies based on the maturity of the technologies along a spectrum. It should feature distinct strategies for already-emergent technologies such as AI, and for emerging technologies such as CRISPR gene editing and quantum computing. It should take an inclusive approach that integrates the perspectives of the commercial sector and civil society from the start, as opposed to the government outlining strategies and values before engaging with other sectors. And it should involve collaboration with international institutions on standards, ethics, and frameworks to ensure that US values governing these technologies align with the values of democracy and the international community, and that US technological capabilities are interoperable with those of its partners and allies.
To begin with, government officials crafting technology strategies must recognize that it is better to understand technologies on a spectrum of maturity based on their current states of R&D and application rather than simply within the binary categories of “emerging” or “not emerging.” Hacking tools and techniques (in a very broad sense) have been around for decades, but they are continually evolving in type and sophistication and now encompass, for instance, escalating attacks on the digital supply chain. In contrast, R&D on quantum computing is still very much ongoing, and its potential applications are only in nascent stages. Physics constraints, resource demands, and the mere temperature needed to keep quantum processors cool, for example, fundamentally differentiate the development of quantum computing from, say, the coding of a particular software exploit. In terms of both R&D and applications, quantum computing is still emerging. Hacking tools and techniques are further along the spectrum of maturity.
Then there is artificial intelligence, which is often lumped into the “emerging technologies” bucket. But modern machine learning techniques have been around for decades, and it was merely advances in computing hardware that enabled so many companies and individuals to put them into practice, resulting in the “deep learning boom” seen in the last two decades. Artificial-intelligence applications are widely deployed already, often in highly troubling ways, across everything from music recommendation and automatic clothing-sizing to healthcare-data analysis, prison sentencing, and facial recognition for police departments. It is thus only in the regulatory sense that artificial intelligence is an “emerging” technology. The US government is only beginning to grapple with how to regulate its development and use, and mistakenly pursuing a generalized strategy for the entire field when there are many uses of artificial intelligence that require just as varied regulation. It is the application of different artificial-intelligence technologies, not just their development, that has bearing on their strategic impact in great-power competition and their potential to be used for good or ill.
The next NDS should also consider a new model of public-private engagement such as a public-purpose consortium, which as described by the physicist Jake Taylor could enable “a competitive approach to research and development before markets exist, while not undercutting progress via duplicated effort or the reduction of scientific transparency.” When the government fails to involve the commercial sector and civil society early enough in the process for developing strategy, it produces a lopsided power dynamic and exacerbates cultural friction and distrust among the three sectors. Establishing a novel form of public-private partnership is especially important given the role that the private sector plays in R&D and that civil society plays in the applications and societal acceptance of emerging technologies. Private-sector players should help ensure that from the outset, government strategies consider challenges with these technologies such as problematic dual-use implications, the vulnerability of algorithms to manipulation, and the insecurity of 5G networks. Civil-society actors should play a key role in advancing the conversation around ethics, standards, and terminology.
That conversation must also occur between the United States and international partners and institutions—another key element to incorporate into the next NDS. Consider, for example, the development and deployment of 5G networks, where the United States is already behind China’s Huawei. The US government needs to focus on building a secure and trusted next-generation technology stack (the layers of hardware, software, and technical protocols that collectively form 5G networks) with NATO members and other allies and partners, many of which have already deployed 5G in their countries. The technology must be able to operate with different telecommunications networks and spectrums across multiple countries and continents, enabling the kind of free, open, and voluntary-standards-based communication that has facilitated the internet’s growth. The US government must also pursue more dialogue and information-sharing with allies and partners on frameworks for a “secure 5G,” and should consider joining similar efforts by multilateral institutions such as the World Bank Group’s Digital Infrastructure Initiative and NATO’s Emerging and Disruptive Technologies roadmap. Discussions of technology ethics, whether they relate to research, development, or applications, should lead to Geneva Conventions-like global norms to ensure that human rights are protected and that technology is used in the public interest.
Digital and emerging technologies, from 5G telecommunications to quantum computing, will transform the environment in which the United States operates and the ways in which it advances national security and global stability. The next NDS must reflect this dawning reality.
Courtesy: The Atlantic Council
