Google X, Deepface, and the Militarization of Consumer Technology

By Ara Anjargolian

research_center

On October 29, 1969, from a lab in UCLA programmer Charley Kline typed the letters “lo” on his keyboard to a counterpart at the Stanford Research Institute. Then his computer crashed. Kline had just sent the first message on ARPANet, a network conceived by the Advanced Research Projects Agency (ARPA), and the precursor to the modern internet.

The trillion dollar lolcat delivery network that the internet would become was certainly not what ARPA intended to create, but it is just one example of military technology transforming the everyday lives of consumers. ARPA also had an early role in developing the predecessor to the Global Positioning System, the technology behind just about every major consumer navigation system on the market today.

Lady_Closet

But ARPANet was powered on in 1969, and the Global Positioning System was largely developed in the 1970s. Which begs the question, what has the military done for consumers lately? Not only have truly transformative technologies largely stopped originating from the military research establishment, but signs of a change in power are increasingly obvious, a change wherein society-transforming technology will originate from consumer-focused technology companies instead of military-focused government agencies and contractors.

Indeed, it seems apparent that the tools of future warfighters and intelligence analysts will have been originally developed to optimize local contextual advertising and automate photo tagging.

To understand why this sea change is occurring, we must go back to 1969 and anti-war U.S. senator Michael Mansfield. A vocal critic of the Vietnam War and fearful for the military industrial complex, Mansfield sought to curtail the military’s influence on society at large. With amendments in 1969 and 1973, Mansfield prohibited military funding for research that lacked a direct or specific military function. Not coincidentally, ARPA renamed itself to DARPA around the same time, tacking the “Defense” in front of its’ name. In subsequent years, the amendments have faced no small amount of criticism. In pushing these amendments, Mansfield inadvertently cut off billions of dollars of future funding for basic research and development that was previously initiated by defense agencies. Looking back to 1950, US spending on research and development, largely driven by defense R&D, has handily outpaced the growth of the US economy.

ycharts_chart (7)

But since 1970, US R&D spending has actually substantially fallen behind the growth of the US economy.

ycharts_chart (8)

It was not Mansfield’s intent to curtail US investment in basic research and development, but it turns out US political will to invest in science and technology is tenuous when not considered within the purview of maintaining military dominance.

Whether the DARPA of 1973 and onwards could have funded ARPANet, claiming it had a “direct or specific” military function, is debatable. However, DARPA in the 1970s and 1980s dutifully followed its new marching orders, investing in research whose military applications are unquestionable. In recent years, the roster of active projects at DARPA reads like a young Darth Vader’s Christmas wishlist. High-energy lasers. Check. Autonomous vehicles and aircraft. Check. Battlefield ready robots and exoskeletons. Check.

Yet the seeds the DARPA and the US industry at large had planted in the early ‘70s really came into fruition starting in the mid to late ‘90s. First the internet exploded in popularity. Ten years later GPS functionality found itself on every phone. On these foundations, the value of technology companies including Google, Facebook, Amazon and Apple skyrocketed. At the same time that these companies were really coming to into their own, the US economy entered the recession so deep even defense spending was not immune. Since 2005 growth in US defense R&D spending looks, unsurprisingly, like a flat line as compared R&D spending growth of some major US consumer technology names.

ycharts_chart (9)

As companies have poured billions of dollars fighting each other to produce the next blockbuster consumer gadget, these masterfully engineered devices have naturally found a military audience. The Economist noticed the phenomenon in 2009 in its article The military-consumer complex noting: “soldiers in Iraq and Afghanistan are using Apple iPods and iPhones to run translation software and calculate bullet trajectories. Xbox video-game controllers have been modified to control reconnaissance robots and drone aircraft.” The Economist article came with what at the time seemed like a prudent disclaimer: “Of course, there are limits to this off-the-shelf approach: it is no way to procure tanks, helicopters or missile systems. But the selective use of existing technology allows military planners to focus their spending on the development of new technologies.” In 2014, even this assertion is up for debate.

Consider Boston Dynamics. Spun off from MIT, the companies spent the better part of the ’90s and ‘00s developing some of the most advanced robots, funded primarily by DARPA. It found its 15 minutes of internet fame in 2009 with a YouTube video of its Big Dog robot (viewed over 16 millions times).



In July 2013, the company debuted it ATLAS robot. The robot will be used by teams to compete in the DARPA Robotics Challenge. On December 14, 2013 Google purchased Boston Dynamics for an undisclosed sum.



The DARPA Robotics Challenge held trials on December 20-21, 2013 to determine a set of team that will compete in the final competition. The hands down winner: SCHAFT, a Japanese robotics company Google had acquired that same year. The New York Times reported that besides SCHAFT and Boston Dynamics, Google has also quietly acquired 6 other robotics companies.

By then, the uneasy links the company had with DARPA were becoming too numerous to name. Another attention garnering Google project that debuted in 2012: the self-driving car, led by engineer Sebastian Thrun. Thrun came to notoriety by leading the team that won the 2005 DARPA Grand Challenge, a race of autonomous vehicles across deserts and mountains near the California/Nevada border.

What about groundbreaking technologies development with no initial involvement of DARPA or the Department of Defense? Google started it’s Google X facility, with the express purpose of taking on “moonshots,” high risk/high reward technological projects and ideas, a mission statement remarkably close to ARPA’s original charter. Its Project Glass program has produced a slim, light head-mounted display that will be sold for less than $1500, raising the prospect that your average early-adopter techie will have better/cheaper “situational awareness” than US troops deployed in Afghanistan.

Yet Google is not alone in its lofty technological aspirations. Facebook employs researchers whose recent publication reveals DeepFace, a new facial recognition algorithm with close to human accuracy. The company also recently acquired Oculus VR and all but promised to write a blank check to fund its further development. Oculus currently boasts the most advanced virtual reality system on the market.

Lesser-known Palantir Technologies develops data analysis software that is by all accounts loved by US intelligence analysts. In the 1960s, the concepts behind the software might have been developed by an ARPA grant to a group of university researchers. Yet Palantir traces its origins to software and methods developed at PayPal to detect and combat payments fraud perpetrated against consumers by Russian crime syndicates.

The point being, an aspiring engineer today who wants to work on computer vision, machine learning, artificial intelligence or advanced robotics is much more likely to end up at Facebook, Google, Palantir, or another deep-pocketed corporation than at a government grant funded university research group. This shift no doubt reflects a great shift in power towards multi-national corporations and away from the governments that regulate them. The cultural and geopolitical implications of a technological arms race paced by the whims and whimsies of consumer electronics and software companies remains to be seen.