A British Advanced Research Projects Agency?

Team_ViGIR_Robot_Florian_DARPA_Robotics_Challenge_2015

Political landscapes across the world are littered with the ruins of “white elephant projects” — expensive, unnecessary, and ultimately useless. The U.K. government has announced it is setting up a British version of the Advanced Research Projects Agency (ARPA, now called DARPA) — the Pentagon unit responsible for huge technological advances during the Cold War. But can they avoid it becoming a failure?

As part of this year’s budget, Chancellor Rishi Sunak announced funding of £800 million ($1.04 billion) to establish the new agency. “We are the country of Newton, Hodgkin, and Turing,” he said during his speech to the House of Commons in March. “Ours is a history filled with ideas, invention, and discovery … To compete and succeed over the next decade and beyond, we need to recapture that spirit.” However, some in the British research community are perplexed. As one Sheffield University professor said, “If you talk to science funders in Washington, they are baffled that Britain is fetishizing something that came out of the Cold War and a very particular context.”

 

 

This new agency is the brainchild of Number 10 advisor Dominic Cummings, a man known throughout Westminster for his iconoclastic approach to government and his eclectic mix of ideas. His love for American innovation culture is clear, from his penchant for wearing Silicon Valley-style hoodies in the office to his recent creation of a Lockheed-esque “skunkworks” unit to drive change within the civil service. A British version of ARPA was his first priority for a post-Brexit United Kingdom — so much so that his WhatsApp status during last year’s election was “Get Brexit done… then ARPA.”

A fresh look at how the United Kingdom supports homegrown technological innovation is welcome. However, Cummings does not seem to fully appreciate the specific context that made ARPA successful, underestimating the defense aspect of the agency while potentially allowing the project to become politicized in a way that makes it unsustainable in the long term. If the new agency is to be successful, Cummings and his colleagues should take the opportunity to reevaluate key parts of this plan.

Lessons from ARPA

ARPA was created by the Eisenhower administration in 1958 within the Department of Defense as a direct response to two instances of technological surprise: the Soviet Union’s successful launch of Sputnik in 1957, and their breach of the nuclear testing moratorium just months later. The U.S. government was caught unprepared, and thus the agency’s singular mission throughout the six decades of its history has been to ensure that the United States became the initiator and not the victim of technological surprise. It makes investments within the wider research ecosystem of governmental, academic, and corporate institutions, and consciously aims for transformational change rather than incremental scientific advances.

In his extensive writing on ARPA, Cummings looks closely at what he sees as its most productive period, from 1962 to 1975. He focuses on the work of one of its departments in particular, the Information Processing Techniques Office, and its head, J. C. R. Licklider. The partnership between ARPA and Xerox’s Palo Alto Research Center under Licklider’s leadership produced a number of ground-breaking innovations, including the first personal computer, and Cummings highlights key lessons to be learned from this. Firstly, he notes that the agency’s high-risk approach to research produced trillions of dollars of value even though many individual projects failed. Secondly, he rightly argues that Licklider’s management style, bringing in great people and not fettering them in bureaucracy, gave the agency the leeway to be creative and led to the emergence of high-performance teams. The lessons that Cummings draws from this period speak to his own post-Brexit priorities for the United Kingdom: promoting scientific research (making Britain “the best place in the world to invent the future”) and boosting government performance by cutting red tape and hiring the best talent.

His position is supported by a recent in-depth report produced by the think tank Policy Exchange (as one might expect given its close relationship to the current government). The report concludes that a British ARPA has great potential to transform technology research and development in the United Kingdom. The report recommends a focus on developing advanced technologies on a 10 to 15-year timescale and advocates the embrace of risk, the empowerment of talented research managers, and the decimation of bureaucracy. As the report’s authors summarize, “The agency should be prepared to fail fast and fail often, with its success judged by the impact of its successes, which should be transformative.”

Where Is Defense?

Given the explicit national security mandate built into ARPA from its creation, made even more obvious by the adding of “Defense” to its name in 1972 and again in 1996, surprisingly little thought seems to have been given to how a British version would work with defense innovation. Cummings appears to either ignore or underestimate the defense implications in his exploration of ARPA’s history, as does the Policy Exchange report.

It is vital not to overlook the importance of the defense context when assessing how and why ARPA worked so well in the period Cummings relies upon for his analysis. Flexible management and a concentrated talent pool clearly contributed to the agency’s success, but he neglects to recognize one of the key underpinnings of this model — the role of the Department of Defense as the funder of a wide range of research and “customer of first resort” for successful projects. ARPA was designed to supplement existing defense research and development by bridging the gap between basic research and commercial applications; the Pentagon facilitated this by providing demand for technologies ahead of demand from the private sector, backed up by a generous budget. While Cummings briefly notes that “military funding in America has often been much more far-seeing and patient than civilian funding,” he does not explicitly address this point about the Pentagon’s key role.

While this position need not necessarily be filled by the defense bureaucracy, it does need to be filled by a government department, and it is unclear whether Cummings has given thought to where this role will fall in the British context if not the Ministry of Defence. The only other obvious candidate within Whitehall (the Department of Business, Energy, and Industrial Strategy) may not have the clout to achieve this; the innovation portfolio has been shunted between civil service departments several times over the last decade alone, and the department itself has gone through considerable reorganizations and mergers (complete with a series of baffling acronyms) over the same period. Some have floated the idea of making the new agency sit within an existing one under the department’s umbrella, such as U.K. Trade and Investment, but placing it deep in an entrenched hierarchy like this would seem to negate the advantages of bureaucratic flexibility Cummings has identified.

A Model for Innovation

The Ministry of Defence has a large budget, an excellent research and development structure, and a great deal of experience in dealing with industry and academia; it would be wasteful not to take advantage of this, or to attempt to replicate it elsewhere for no good reason.

The research arms of the various parts of the armed forces were consolidated in 1995 into the Defence Evaluation and Research Agency, which sat within the Ministry of Defence and became the country’s largest science and technology research organization. This agency was split into two in 2001. The majority part became a commercial firm, QinetiQ, which was floated on the stock exchange in 2006 with the Ministry of Defence retaining a “golden share”; the firm has since acquired a number of smaller defense contractors and does a great deal of work internationally, including as a trusted supplier to the U.S. government. The remainder of the agency became the Defence Science and Technology Laboratory, an executive agency of the Ministry of Defence. It performs functions that were deemed to be better done by government, including nuclear, chemical, and biological research at the Porton Down laboratory, where the Novichok used in the poisoning of Sergei and Yulia Skripal in Salisbury was identified.

The Defence Science and Technology Laboratory is the primary channel through which the Ministry of Defence engages with industry and academia, running a range of competitions and procurement activities in the research and development space. The department also maintain its relationships with the larger contractors through the Defence Suppliers Forum, which has a subgroup focusing on innovation. There is also the Defence and Security Accelerator, a cross-government team that finds and funds exploitable defense innovation projects through themed competitions.

Recent governments have explicitly recognized the importance of defense innovation. Following the 2018 Modernising Defence Programme consultation, the Ministry of Defence published the Defence Innovation Priorities white paper in late 2019, coupled with the Defence Technology Framework. These documents show the British government’s laudable determination to grasp the opportunities of new technologies in the defense arena, acknowledging that the nations that are best able to exploit technological advances may gain a decisive edge in future conflicts. There is an explicit recognition of the need to better bring together the United Kingdom’s governmental, academic, and industrial research capabilities to further the advancement of defense technologies — and this will no doubt be a big part of the forthcoming Defence Review, which has been delayed by the COVID-19 pandemic. Like the Pentagon for ARPA, the Ministry of Defence is best placed to support the new agency, and Cummings should consider this more closely.

Can ARPA Be Replicated At All?

The seductive simplicity of ARPA’s structural tenets, combined with its success at boosting innovation, have led other areas of the American federal government to attempt to replicate it. Two of these stand out, both established during the second term of the George W. Bush administration: the Department of Energy’s Advanced Research Projects Agency – Energy and the Office of the Director of National Intelligence’s Intelligence Advanced Research Projects Agency. Each is oriented to its specific sector, but both benefitted in their early years from the expertise of those closely familiar with the ARPA mindset. While their existence may yet be too short to critically assess their success in fostering significant technology transitions, there are some encouraging signs. A recent National Academies report found that the Department of Energy’s agency has made significant progress towards its goal of fostering breakthrough energy technologies. While the intelligence version focuses on its agency customers rather than on commercialization, almost three quarters of its developed projects have achieved at least one transition to the intelligence sector. However, the question remains as to whether these achievements could have been made without the establishment of an ARPA-like structure — the Department of Energy, for example, has an existing network of laboratories that could well have created similar results if they had been given that funding.

It is also important to recognize that the technology base has changed significantly since ARPA’s founding, and since its period of heightened success in the 1960s and early 1970s. The primary driver at the time was “spin-off” — the federally funded laboratories did the research and created the military advances, and the commercial sector took up those technologies that could be translated into the commercial sphere. Commonplace civilian technologies like the microwave, the navigation systems in our cars, and the internet itself all have their roots in defense research and development that was funded by the U.S. government. This kind of context fits well with the ARPA model, which did a great deal to fund basic research and to facilitate the spin-off technology transitions that emerged as a result. However, today’s context has moved significantly towards “spin-on” — breakthrough technologies are now just as likely (if not more likely, in some sectors) to be created in the commercial sphere, and private firms are seeing the benefits of investing in their own research and development arms instead of waiting for manna to fall from the government. Innovation is increasingly global and commercial, and militaries all over the world (including those of the United States and the United Kingdom) have an appreciation of the increasing salience of commercial off-the-shelf technologies in the procurement mix. It is simply not clear that the advantages brought by the early days of ARPA can be recreated today given the far smaller role that government research and development of any kind now has in the technology base.

Dangerous Politicization

Innovation, particularly of the high-risk ARPA variety, involves frequent and repeated failure — not something that governments find easy to weather. While private venture capital firms can build this into their long-term investment models, democratically elected governments face the prospect of explaining failure to voters on a regular basis in a context where their opposition will seize on any prospect of an expensive “white elephant” to shout about. One can easily imagine a general election in 2025 where the Conservative party is castigated for spending hundreds of millions of pounds on an agency that hasn’t produced an obvious success, despite the fact that four or five years is simply not long enough to effectively judge the record of an agency of this type. A change of governing party may also see the agency defunded entirely; an incoming Labour government with no buy-in or attachment to a British ARPA could simply abolish it.

A successful innovation project on an ARPA model needs decades of sustained investment, likely from a succession of governments of different political stripes. If a British ARPA is to thrive, it therefore needs to be depoliticized, but Cummings’s clear leadership of the project does not make this an easy task. He is a polarizing figure at the best of times, and is extremely close to the prime minister, Boris Johnson; many on both sides of the political spectrum view him as the power behind the throne. Following his successes in campaigning for Brexit, Cummings has chosen the overhaul of the Whitehall civil service as his new project and has expressly linked the ARPA idea to this wider plan of tackling the fossilized bureaucracy of the British state. If all of his projects within this context are so explicitly linked, there is the risk of failure elsewhere bringing the whole edifice crashing down onto the new agency. While a fresh look at British innovation is a good idea, and deeper thought on this subject is always welcome, this part of the plan may succeed more readily if Cummings is willing to let go of it and allow a wider consensus to emerge.

Reevaluation Needed

ARPA provides a clear model for a research agency based on creative freedom, technology outcomes, and reaching for breakthrough advances, and Cummings is right to highlight it as a good example to draw upon when considering how best to improve innovation in the United Kingdom. However, it cannot be divorced from its context, both in terms of the time period during which it was most successful and its overt placement within the defense field. While the early history of ARPA makes for appealing reading, that success cannot be replicated simply through fetishizing Licklider’s approach. It is entirely possible to improve innovation in the United Kingdom (or, indeed, in any nation) by making the interconnection of government, industry, and academia run more smoothly, but it is by no means clear that the achievement of this goal requires the creation of an ARPA-like agency.

If a British version of ARPA must be established, it must indeed be truly British: appropriate to the current context, depoliticized, and designed to fit within the existing innovation ecosystem and technology base. Any new research agency should take advantage of the expertise found within the Ministry of Defence, and this appears to be a serious blind spot in Cummings’s plans. The government should be building on what the United Kingdom does well, not simply transplanting a relic of the early Cold War from across the Atlantic. Hopefully these issues will be addressed as the process matures, and the looming prospect of an expensive “white elephant” will be avoided.

 

 

Emma Salisbury is working on her Ph.D. at Birkbeck College, University of London. Her research focuses on defense research and development in the United States and the military-industrial complex. She is also a senior staffer at the U.K. Parliament. The views expressed here are solely her own. You can find her on Twitter @salisbot.

Image: Wikicommons (Photo by Sun L. Vega)