Photo: jlee31180 (Flickr).

In September 2011, the CIA and the U.S. military jointly launched a drone strike authorized by President Barack Obama. The attack resulted in the killing in Yemen of Anwar al-Awlaki, a fervent Muslim cleric born in the United States. The organizers of the attack relied on Awlaki's geolocation data, controlled by the National Security Agency as part of a surveillance program. Two weeks later, a CIA drone strike killed another U.S. citizen using the same type of data: al Awlaki's 16-year-old son, Abdulrahman al Awlaki.

Although al-Awlaki was deliberately killed by U.S. forces, other U.S. citizens — and thousands of civilians in Afghanistan and elsewhere in Central Asia and the Middle East — have been inadvertently killed by drones. These cases portend a major flaw in the latest iteration of automated warfare: the imprecision of technologies and the large margins of error that accompany even the most sophisticated new weapons systems. In their most advanced form, computerized tools make use of artificial intelligence and machine learning, and could soon have fully autonomous capabilities.

Internet-connected portable digital devices have transformed billions of people around the world into atomized data-producing machines, which input information into hundreds, if not thousands, of algorithms every day. Although we have rapidly integrated smartphones and tablets into our lives, we rarely reflect on how the data stored and transmitted by these devices can be easily militarized. For example, recent reports describe how the U.S. Defense Intelligence Agency, affiliated with the Department of Defense (DoD), routinely uses commercially available geolocation data collected from individual mobile phones, sometimes without a warrant. Military and intelligence agencies can use that data not only to spy, but also to reconstruct social networks and even to direct lethal attacks against individuals.

Drones, geolocation software, spyware and other such tools are emblematic of a new series of collaborations between Big Tech and Big Defense. For the past two decades, the Department of Defense and 17 U.S. government agencies collectively known as the U.S. Intelligence Community have tried to capture technological innovation at its source: Silicon Valley. Military and spy agencies have done so by creating outposts along the West Coast; organizing a high-level advisory council linking the Pentagon to big tech companies; coordinating summits, forums and private meetings with influential investors and business executives; and appealing directly to the hearts and minds of entrepreneurs, engineers, computer scientists and researchers who are sometimes skeptical of government bureaucrats, especially those in the DoD.

In many ways, it's impossible to fully understand the U.S. military today without an analysis of its deep connections to the tech industry.

"Internet-connected portable digital devices have transformed billions of people around the world into atomized data-producing machines, which feed information into hundreds, if not thousands, of algorithms every day."

The interconnections between the worlds of network technology and defense date back more than 50 years. For example, from the early 1960s, the Department of Defense's Advanced Research Projects Agency (ARPA) played a crucial role in funding the computer research that gave rise to the ARPANET, the precursor to today's Internet. The initial development of Silicon Valley was largely funded by defense and intelligence agencies, and the Pentagon invested heavily in tech companies (external link) throughout the Cold War.

What is virtual warfare?

Obviously, virtual warfare means different things to each person. There is no agreed definition, which allows the term to be interpreted broadly, holistically and anthropologically. I take a broad perspective, focusing on four different elements: robotic and autonomous weapons systems; a high-tech version of psychological operations or psyops; predictive modelling and simulation software, referred to by some as "computational counterinsurgency"; and cyberwarfare, that is, the attack and defense of critical infrastructure. These technologies and techniques are based on the production, availability, and analysis of massive amounts of data—often surveillance data—collected from drones, satellites, cameras, mobile phones, electronic transactions, social media, emails, and other Internet sources.

We can think of this as an algorithm war. Increasingly, technologies make use of artificial intelligence or AI to automate decision-making processes. The development of virtual weapons depends on the combined efforts of a wide range of scientists and technical experts: not only chemists, physicists, engineers, computer programmers and data analysts, but also biotechnology researchers, political scientists, psychologists and anthropologists. Much of the work is quite banal and takes place in nondescript buildings of suburban office parks, technology campuses or university laboratories. Silicon Valley has become one of the main centers of this type of defense and intelligence work.

In a way, virtual warfare is a continuation of the so-called Revolution in Military Affairs or RMA, a doctrine articulated by the Pentagon's Network Assessment Office in the eighties and nineties. It leaned heavily towards technology-based solutions. After September 11, when the United States launched its so-called Global War on Terror, and went to war against global networks of insurgents armed with relatively simple technologies such as improvised bombs, rifles and grenade launchers, the RMA lost steam, and counterinsurgency became fashionable after a long hiatus. But now, in a period marked by rapid innovation, algorithmic modes of governance, and the rise to power of rival nations like China and Russia—each pursuing its own virtual warfighting technologies—computerized combat has once again taken center stage among the elites of the U.S. military establishment.

The Intersection of Big Defense and Big Tech: Creating an IDIx

Mountain View rests comfortably between the forested Santa Cruz Mountains and the south shore of San Francisco Bay. During the first half of the twentieth century it was a quiet town with cattle farms, fruit orchards and picturesque downtown streets. But after a team of scientists led by William Shockley invented the semiconductor there in 1956, it grew rapidly, along with the rest of Silicon Valley. Today it is a bustling suburb with more than 80,000 residents.

At first glance, it seems like a strange place for military and intelligence agencies to be established. Mountain View is almost 4,024 km from the Pentagon. Direct flights from San Francisco to Honolulu take less time than flights to Washington, DC.

Not only are the Pentagon and Silicon Valley geographically far apart, but there are other differences as well. The Department of Defense is often seen as a notoriously bloated, stretched, and wasteful bureaucracy, with rigidly hierarchical organizational structures and inflexible labor standards. Instead, Mountain View's largest employer is Alphabet, the parent company of Google, one of the world's most valuable corporations. Its 26-acre campus, known as the Googleplex, includes more than 30 cafeterias, free food and drink for its employees, gyms and swimming pools. A life-size iron Tyrannosaurus rex skeleton, affectionately called Stan by Google employees, features prominently on the exterior of a main building.

Despite these differences—indeed, because of them—Defense Secretary Ash Carter very publicly established a Pentagon outpost less than three miles from the Googleplex. The Experimental Defense Innovation Unit, or DIUx, was created in August 2015 to quickly identify and invest in companies developing cutting-edge technologies that could be useful to the military. With DIUx, the Pentagon created its own startup accelerator dedicated to funding companies specializing in AI, robotic systems, big data analytics, cybersecurity and biotechnology.

The new home of IUDs wasn't so out of place. Its headquarters were in a building once occupied by the Army National Guard, on the grounds of the Ames Research Center, the largest of NASA's ten field centers, and Moffett Field, once home to the 130th Rescue Squadron of the California Air National Guard. Defense giants Lockheed Martin and Northrop Grumman have offices less than 3 km away. In 2008, Google itself invaded government territory: it signed a 40-year lease with NASA Ames for a new research campus. He then signed a sixty-year agreement with NASA to lease 1,000 acres of Moffett Field, including three huge airship hangars. Today, Google uses hangars to build stratospheric balloons that could one day provide Internet services to rural dwellers, or perhaps carry out high-altitude military surveillance missions.

"With DIUx, the Pentagon created its own startup accelerator dedicated to funding companies specializing in AI, robotic systems, big data analytics, cybersecurity and biotechnology."

The DIUx office was very close to other tech companies: Amazon's Lab126 (where the Kindle reader, Amazon Echo and other digital devices were created); LinkedIn's corporate headquarters; and Microsoft's Silicon Valley campus. Apple's corporate offices were 8 km away, in nearby Cupertino. The Pentagon's new facility was literally at the intersection of Big Tech and Big Defense. The DIUx office, located in a brick building, embraced the contradictions of the Pentagon West: "The corridors are monotonous, the doors have combination locks. But inside, the newcomers have renovated the spaces with whiteboards, whiteboards and desks arranged in random diagonals, matching the non-hierarchical atmosphere of a Valley startup," reported an observer.


Ash Carter's plan was ambitious: harnessing the best and brightest minds in the tech industry for use by the Pentagon. This Pennsylvania native had spent several years at Stanford University prior to his appointment as secretary of defense, and he was impressed with the innovative spirit and billionaire tycoons of the Bay Area: They are inventing new technologies, creating prosperity, connectivity and freedom," Carter said. "They feel that they, too, are public servants and would like to have someone in Washington with whom they can connect." Remarkably, Carter was the first sitting defense secretary to visit Silicon Valley in more than 20 years.

The Pentagon has its own research and development (R+D) agency, DARPA, but it pursues projects that are decades away, not months. Carter wanted a nimble, streamlined office that could act as a sort of go-between, funneling tens or hundreds of millions of dollars from the Defense Department's huge budget to promising companies developing nearing completion technologies. Ideally, DIUx would serve as a liaison, negotiating the needs of four-star generals, Pentagon civilian leaders, and hooded engineers and businessmen. Soon, DIUx opened branches in two other cities with burgeoning tech sectors: Boston and Austin.

In the short term, Carter hoped DIUx would establish relationships with local startups, recruit top talent, involve military reservists in projects, and streamline the Pentagon's notoriously cumbersome procurement processes. His long-term goals were even more ambitious: to take career military officers and assign them to work on futuristic projects in Silicon Valley for months, to "expose them to new cultures and ideas that can take them back to the Pentagon . . . [and] invite coaches to spend time in Defense."

In March 2016, Carter organized the Defense Innovation Council (DIB), an elite civilian think tank tasked with advising and making recommendations to Pentagon leadership. He appointed former Google CEO and Alphabet board member Eric Schmidt to chair the DIB, which included current and former executives from Facebook, Google and Instagram, among others.

Three years after Carter launched DIUx, it was renamed the Defense Innovation Unit (DIU), indicating that it was no longer experimental. Despite initial difficulties, Deputy Defense Secretary Patrick Shanahan described DIUx as "a valuable and proven asset." "The organization itself is no longer an experiment," he said in 2018. "IUDs remain vital to fostering innovation across the Department and transforming the way the DoD builds a more lethal force." In early 2018, the Trump administration requested a sharp increase in the IUD budget for fiscal year 2019, from $30 million to $71 million. For 2020, the administration requested $164 million, more than double the previous year's request.

"In March 2016, Carter organized the Defense Innovation Council (DIB), an elite civilian think tank tasked with advising and making recommendations to Pentagon leadership."

The CIA's venture capital fund

Although Pentagon officials presented DIUx as an innovative organization, it was actually modeled after another company created to provide a similar service to the U.S. intelligence community. In the late 1990s, the CIA created a non-profit entity called Peleus to capitalize on innovations being developed in the private sector, with a focus on Silicon Valley. Soon after, the organization was renamed In-Q-Tel.

The first CEO, Gilman Louie, described how the organization was created to solve "the big data problem":

[CIA leaders] were very afraid of what they called at the time the prospect of a 'digital Pearl Harbor.' Pearl Harbor happened when every part of the government had a piece of information, but they couldn't put it together to say, "Look, the attack on Pearl Harbor is imminent." In 1998, they began to realize that information was isolated across all these intelligence agencies and that they could never put it together. They were trying to solve the big data problem. How to join them to get information?

By funneling CIA funds to nascent companies creating surveillance, intelligence gathering, data analysis and cyberwarfare technologies, the agency hoped to gain an edge over its global rivals by co-opting engineers, hackers, scientists and creative programmers. In 2005, the CIA invested approximately $37 million in In-Q-Tel. By 2014, the organization's annual funding had skyrocketed to nearly $94 million and it had made 325 investments in a staggering array of tech companies.

If In-Q-Tel looks like something out of a James Bond movie, it's because the organization was inspired in part by the Q branch, the fictional R+D office of the British secret service, popularized in Ian Fleming's spy novels and Hollywood blockbusters. In-Q-Tel and DIUx were ostensibly created to transfer emerging technologies from the private sector to U.S. intelligence services and military agencies, respectively. A somewhat different interpretation is that these organizations were created "to capture technological innovations... [and] to capture new ideas." Critics point to In-Q-Tel as a paradigmatic example of the militarization of the tech industry.

In monetary and technological terms, In-Q-Tel's most profitable investment was probably Keyhole, a San Francisco-based company that developed software to weave together satellite imagery and aerial photos to create three-dimensional models of the Earth's surface. The program could essentially create a high-resolution map of the entire planet. In-Q-Tel provided funding in 2003 and, within months, the U.S. military used Keyhole to support U.S. troops in Iraq.

“Si In-Q-Tel parece sacada de una película de James Bond, es porque la organización se inspiró en parte en la rama Q, la oficina ficticia de I+D del servicio secreto británico, popularizada en las novelas de espionaje de Ian Fleming y en las superproducciones de Hollywood”.

Las fuentes oficiales nunca revelaron cuánto invirtió In-Q-Tel en Keyhole, pero en 2004 Google compró la start-up. Pasó a llamarse Google Earth. La adquisición fue significativa: el periodista Yasha Levine escribe que el acuerdo Keyhole-Google “marcó el momento en que la empresa dejó de ser una compañía de Internet puramente orientada al consumidor y comenzó a integrarse con el Gobierno de Estados Unidos“. En 2007, Google buscaba activamente contratos gubernamentales repartidos uniformemente entre agencias militares, de inteligencia y civiles.

Además de Google, la cartera de In-Q-Tel incluye empresas con proyectos futuristas como Cyphy, que fabrica drones anclados que pueden volar en misiones de reconocimiento durante largos periodos gracias a una fuente de energía continua; Atlas Wearables, que fabrica rastreadores de actividad física que controlan de cerca los movimientos corporales y las constantes vitales; Fuel3d, que vende un dispositivo portátil que produce escaneados tridimensionales detallados de estructuras u objetos; y Sonitus, que ha desarrollado un sistema de comunicaciones inalámbricas, parte del cual cabe dentro de la boca del usuario. Si DIUx ha apostado por empresas de robótica e inteligencia artificial, In-Q-Tel ha perseguido a las que crean tecnologías de vigilancia: empresas de satélites geoespaciales, sensores avanzados, equipos de biometría, analistas de ADN, dispositivos de traducción de idiomas y sistemas de ciberdefensa.

Más recientemente, In-Q-Tel se ha orientado hacia empresas especializadas en la extracción de datos de redes sociales y otras plataformas de Internet. Entre ellas figuran Dataminr, que transmite datos de Twitter para detectar tendencias y posibles amenazas; Geofeedia, que recopila mensajes de redes sociales indexados geográficamente en relación con noticias de última hora, como protestas; y TransVoyant, una empresa que recopila datos de satélites, radares, drones y otros sensores.

Algunos podrían aplaudir el éxito de la contratación de empresas tecnológicas por parte del ejército y las agencias de inteligencia estadounidenses. Dado el rápido desarrollo y despliegue de sistemas de armamento y programas de vigilancia de alta tecnología por parte de naciones rivales como China –que ha desplegado tecnologías comparables contra sus propios ciudadanos en la provincia de Xinjiang—, sus defensores suelen afirmar que el ejército estadounidense no puede permitirse quedarse rezagado en una carrera armamentística de IA. Pero estos argumentos no tienen en cuenta cómo la fusión de la Gran Defensa con otra gran industria vinculará aún más estrechamente la economía estadounidense a guerras interminables en el extranjero y a la militarización de la policía en el país.

Proyecto Maven

Muchas de las empresas financiadas por In-Q-Tel y DIUx han sido pequeñas start-ups muy necesitadas de liquidez. Pero el interés del Pentágono por Silicon Valley también se extiende a las mayores empresas basadas en Internet.

Consideremos el caso del Proyecto Maven, conocido formalmente como el Equipo Interfuncional de Guerra Algorítmica. El subsecretario de Defensa Robert Work estableció el programa en abril de 2017, describiéndolo como un esfuerzo “para acelerar la integración del DoD de big data y aprendizaje automático. … [y] convertir el enorme volumen de datos de que dispone el DoD en inteligencia y conocimientos procesables a gran velocidad…”. El Bulletin of the Atomic Scientists expone el problema de forma sucinta:

Los aviones espía y los satélites estadounidenses recogen más datos en bruto de los que el Departamento de Defensa podría analizar aunque toda su plantilla se dedicara a ello durante toda su vida. Por desgracia, la mayor parte del análisis de imágenes implica un trabajo tedioso: la gente mira las pantallas para contar coches, personas o actividades. …la mayoría de los datos de los sensores desaparecen, nunca se analizan, a pesar de que el Departamento lleva años contratando analistas a toda velocidad.

El Pentágono había gastado decenas de miles de millones de dólares en sensores. Crear algoritmos para clasificar y analizar las imágenes tenía sentido desde el punto de vista económico, y con un coste previsto de 70 millones de dólares, el Proyecto Maven debía de parecer una ganga. El alcance del trabajo era asombroso. En su estado actual, los sistemas de IA requieren conjuntos de datos masivos para el “aprendizaje profundo”, que esencialmente significa aprender con el ejemplo. Durante la segunda mitad de 2017, las personas que trabajaban en el Proyecto Maven etiquetaron más de 150.000 imágenes visuales para crear los primeros conjuntos de datos para entrenar los algoritmos. Las imágenes —fotos de vehículos, personas, objetos, eventos— tenían que tener en cuenta cientos, si no miles, de condiciones variables: diferentes altitudes, ángulos de foto, resolución de imagen, condiciones de iluminación, etc.

¿Qué organización podría encargarse de semejante tarea? Los funcionarios del Pentágono no dijeron nada sobre las empresas implicadas, pero algunos informadores dieron pistas indirectas de la participación de importantes empresas tecnológicas. El coronel del Cuerpo de Marines Drew Cukor que dirigió el Proyecto Maven, señaló que “estamos en una carrera armamentística de IA. Está ocurriendo en la industria [y] las cinco grandes empresas de Internet lo están persiguiendo con ahínco”. Muchos de ustedes habrán notado que Eric Schmidt [entonces consejero delegado de Alphabet Inc, la empresa matriz de Google] llama ahora a Google una empresa de IA, no una empresa de datos”.

Apenas ocho meses después de que el Departamento de Defensa pusiera en marcha el Proyecto Maven, los militares utilizaban los algoritmos del programa para apoyar las misiones con drones contra el ISIS en Irak y Siria.


En marzo de 2018, Gizmodo publicó una serie de ampollas revelando que el Pentágono había contratado silenciosamente a Google para el trabajo del Proyecto Maven en septiembre de 2017. Según correos electrónicos internos de ejecutivos de Google, el acuerdo estaba valorado en al menos 15 millones de dólares, y se esperaba que aumentara hasta 250 millones.

Algunos correos electrónicos detallaban reuniones entre ejecutivos de Google y el subsecretario de Defensa Jack Shanahan. Más de diez empleados de Google estaban asignados al proyecto, y la empresa se había asociado con varias otras empresas, entre ellas DigitalGlobe, una empresa de imágenes geoespaciales, y CrowdFlower, una empresa de crowdsourcing. CrowdFlower (que desde entonces ha cambiado su nombre por Figure Eight) pagó a los llamados “crowd workers” —personas que realizan tareas repetitivas en línea, como identificar fotos— para que etiquetaran miles de imágenes para el “aprendizaje profundo” algorítmico. Al parecer, los trabajadores no sabían lo que estaban construyendo ni quién se beneficiaría de ello.

Algunos de los correos electrónicos internos de Google daban a entender que la empresa tenía planes ambiciosos que iban más allá de lo sugerido en los anuncios iniciales del Pentágono. Uno de ellos sugería la creación de un sistema de espionaje “similar a Google Earth” que permitiera a los usuarios “hacer clic en un edificio y ver todo lo relacionado con él”, incluidas personas y vehículos.

Los responsables de Google se preocupaban en privado por un posible problema de relaciones públicas si se filtraba el proyecto Maven: “Creo que deberíamos hacer buenas relaciones públicas sobre la colaboración del Departamento de Defensa con GCP desde el punto de vista de la tecnología en la nube (almacenamiento, red, seguridad, etc.)”, escribió Fei-Fei Li, científico jefe de IA de Google Cloud, “pero evitar a toda costa cualquier mención o implicación de la IA”.

Pero al final se corrió la voz.

Revuelta de los ingenieros

En febrero de 2018, los correos electrónicos internos sobre el Proyecto Maven circularon ampliamente entre los empleados de Google, muchos de los cuales estaban conmocionados y consternados por lo que habían hecho los altos ejecutivos de la compañía. En cuestión de meses, más de 4.000 investigadores de Google habían firmado una carta dirigida al CEO Sundar Pichai, exigiendo la cancelación del contrato de Maven. La carta, firmada por varios ingenieros de alto nivel, comenzaba así: “Creemos que Google no debería estar en el negocio de la guerra”. También exigía que Google desarrollara “una política clara que establezca que ni Google ni sus contratistas desarrollarán nunca tecnología bélica”. A finales de año, casi una docena de empleados dimitieron en protesta por los contratos militares de la empresa y la falta de transparencia de sus directivos.

Sorprendentemente, los empleados tuvieron éxito, al menos momentáneamente. A principios de junio, Google anunció que la empresa pondría fin a su trabajo en el Proyecto Maven cuando expirara el contrato. Días después, Google publicó una serie de directrices éticas o “principios de la IA“, en las que afirmaba que la empresa “no diseñará ni desplegará IA” para sistemas de armamento, para “vigilancia que viole normas internacionalmente aceptadas”, ni para tecnologías utilizadas para contravenir el derecho internacional y los derechos humanos.

El compromiso de Google de cancelar su trabajo en el Proyecto Maven era demasiado bueno para ser cierto. En marzo de 2019, The Intercept obtuvo un correo electrónico interno de Google en el que se indicaba que una empresa externa seguiría trabajando en el Proyecto Maven utilizando “Google Cloud Platform off-the-shelf (servicio básico de computación, en lugar de Cloud AI u otros servicios en la nube) para soportar algunas cargas de trabajo”. Walker añadió que Google estaba trabajando con “el Departamento de Defensa para realizar la transición de forma coherente con nuestros principios de IA y compromisos contractuales”.

Otros informes revelaron que el Departamento de Defensa había adjudicado el contrato del Proyecto Maven a Anduril Industries, más conocida por crear el casco de realidad virtual Oculus Rift. El año anterior, Anduril había puesto a prueba un sistema de vigilancia desarrollado para los agentes de Aduanas y Protección de Fronteras de Estados Unidos. El sistema utiliza IA para detectar la presencia de personas que intentan cruzar la frontera estadounidense.

Although the media implied that Google (and later Anduril) were the only companies playing a role in Project Maven, the reality is far more complex and troubling. A thorough analysis by the non-profit research organization Tech Inquiry documents the deep involvement of numerous other contractors and subcontractors. The Pentagon awarded "principal awards" to ECS Federal and Booz Allen Hamilton, and "subcontractors" to a number of companies including Microsoft, Clarifai, Rebellion Defense, Cubic Corporation, GATR Technologies, Technical Intelligence Solutions and SAP National Security Services, among others. These contracts were never made public.

Although Google employees who opposed Project Maven represented only a modest portion of the company's 70,000 employees, they managed to spark a debate about tech industry military contracts and a broader debate about AI ethics.

Google's revolt resonated throughout Big Tech and inspired others to follow. For example, in February 2019, more than 200 Microsoft employees demanded that the company cancel a $480 million contract with the U.S. Army to supply troops with more than 100,000 HoloLens augmented reality headsets. The Pentagon's request for proposals raised the need for a head-mounted display capable of providing soldiers with night vision, aiming stealthily weapons and automatically recognizing threats. Ideally, it should provide soldiers with "increased lethality, mobility and situational awareness," according to the announcement.

In an open letter to Microsoft CEO Satya Nadella, workers expressed concern that, in the hands of the military, HoloLens could be "designed to help kill," "turning war into a simulated video game." The employees added: "We do not sign up to develop weapons, and we demand to have a say in how our work is used." Microsoft executives refused to cancel the contract. Nadella said, "We're not going to withhold technology from the institutions we've chosen in democracies to protect the freedoms we enjoy."

During the summer of 2018, approximately 450 employees of tech giant Amazon signed a letter demanding that the company stop selling Rekognition — a facial recognition software program — to law enforcement. The employees' letter also asked that Amazon's Web Services division stop hosting Palantir, a technology company that supplied data analysis software to U.S. Immigration and Customs Enforcement, as it was deporting unaccompanied minors and their families. Amazon CEO Jeff Bezos shrugged at the employee letter. "One of the senior leadership team's jobs is to make the right decision even when it's not popular," he said in October 2018. "If Big Tech is going to turn its back on the U.S. Department of Defense, this country is going to be in trouble."

"During the summer of 2018, approximately 450 employees of the tech giant Amazon signed a letter demanding that the company stop selling Rekognition — a facial recognition software program — to law enforcement."

While tech workers expressed reluctance to get involved in military projects, executives sold their companies' wares to Pentagon officials. Microsoft announced Azure Government Secret, a cloud service for Department of Defense and intelligence community customers who require "US Secret-classified workloads." Oracle's websites boasted about how its products "help military organizations improve efficiency, mission readiness, and execution." And Amazon created a slick ninety-second promo video in August 2018, titled simply 'Amazon Web Services for the Warfighter.'


Against the fusion of big tech and big defense

Silicon Valley technologies illustrate the unforeseeable consequences of the release of new hardware or software. The idea that an invention can be used for peaceful or military purposes — that is, the notion of dual-use technology — has become widespread in American society over the past century. Historian Margaret O'Mara reminds us that during the Cold War, "the Valley built small: microwaves and radars for high-frequency communications, transistors and integrated circuits. Silicon Valley built sleek miniaturized machines that could power missiles and rockets, but also had possibilities for peaceful use: watches, calculators, appliances and computers."

These technologies continue to have dual-use applications. Google Earth can be used for mapping and geographic research, but it can also be used by Special Forces teams to attack power grids, bridges, or other infrastructure. Microsoft first marketed the HoloLens as an augmented reality device for gamers, artists and architects, but the most profitable consumers are likely to be foot soldiers. Amazon's facial recognition program could be used for banking or secure ATM transactions, but they can also be used as surveillance technologies by military, intelligence or law enforcement agencies such as U.S. Immigration and Customs Enforcement. Cloud platforms offered by Amazon, Oracle, Microsoft and Google can store data for scientific researchers, public health officials or commercial companies. But they can also increase the lethality of military forces.

Some might dismiss Google's dissident engineers and scientists as naïve. After all, they didn't know what they were getting into? If scientists usually understand that once they produce knowledge, they probably won't have control over how it's used, then surely they should understand that the devices and apps they were creating could at some point become weapons. Or not?

Many scientists and engineers who now oppose Silicon Valley's military work may never have imagined that they would be dragged into the military-industrial-technological complex. Maybe they even decided to work for tech companies because they thought those companies weren't in the arms business. After all, the letter written by Microsoft protesters states, "We do not sign on to develop weapons."

It's also possible that the researchers placed inordinate faith in their company's executives. At Google, employees felt betrayed by the secret decisions that led to the Project Maven contract. The business press regularly acknowledges that the company has the best "corporate culture" in the United States, not only because employees can bring pets to work and have access to organic meals prepared by professional chefs, but also because the organization has a reputation for valuing employee collaboration.

Once Project Maven came to light, the false consciousness of tech workers began to evaporate. Earning six figures as an engineer or programmer fresh out of college makes it hard to think of yourself as a proletarian, especially when you enjoy the advantages the industry offers: free gourmet lunches, free gyms and nurseries, for example. For thousands of employees, being excluded from debates about whether the company should collaborate on the development of artificial intelligence weapons sparked a latent sense of class consciousness.

There was also another problem: Silicon Valley's longstanding ties to the Pentagon. As explained in this essay and as Margaret O'Mara points out, "Whether your employees realize it or not, all of today's tech giants contain some defense industry DNA..." This implies a much fuller recognition of the long and complicated history of Silicon Valley and the business of war."

"Earning six figures as an engineer or programmer fresh out of college makes it difficult to think of yourself as a proletarian, especially when enjoying the advantages offered by industry"


The divide between the Pentagon and Silicon Valley is mostly a myth: it has never really existed, at least in a meaningful way. The differences are superficial and stylistic. For nearly a century, the regional economy and culture have been determined by what might be called the military-industrial-university complex. During the Cold War, the Pentagon helped build the computer industry by awarding military contracts in fields such as microwave electronics, missile and satellite production, and semiconductor research.

Historian Thomas Heinrich reminds us that popular descriptions of "ingenious inventors-entrepreneurs and venture capitalists [who] forged a dynamic high-tech economy free from the heavy hand of government" divert attention from the crucial role of "Pentagon funding for research and development [that] helped lay the technological foundation for a new generation of startups" in the twenty-first century. From the '50s to the late '90s, Silicon Valley's largest private-sector employer wasn't Hewlett Packard, Apple, Ampex, or Atari. It was the defense giant Lockheed. Today the region faces a familiar pattern, although the size and influence of today's technology companies dwarf the computer companies of yesteryear.

This is likely to have important implications in the near future. Jack Poulson, former research scientist at Google and co-founder of Tech Inquiry

This is likely to have important implications in the near future. Jack Poulson, a former research scientist at Google and co-founder of Tech Inquiry (external link), explained it to me like this: "I think we're witnessing the transition of major U.S. tech companies to defense contracts and I would venture to predict that in the next few years they will buy defense contractors, sort of like Amazon buying Raytheon."

The real dividing line is not between the Pentagon and Silicon Valley. It's inside Silicon Valley, where a modest contingent of politically awakened engineers and scientists have opposed the militarization of their work. When faced with an all-out assault of PR messages, awareness campaigns, "collaborative" debates, more compensation and perks — and perhaps the unspoken threat of losing their jobs or being outsourced — will they give up?

Right now it's too early to know the outcome, but the future of virtual warfare and digital battlefields may be in your hands.

(Taken from ALAI)