News | War / Peace - USA / Canada When the Pentagon Comes to Silicon Valley

Pulling back the veil on Big Tech’s missing moral compass


Photo: Wikipedia

Imagine being instructed by your employer to build a product that will be used to wage war, even end the life of another. Imagine its purpose was only revealed to you after you contributed to its creation. Had you known, would you still have agreed to be involved? What would you do if you were asked not to speak about the situation to anyone outside the company?

Would you, an ordinary civilian, still want to work for a company that helps facilitate war, the suffering of others, and death? This is the situation hundreds of employees of Silicon Valley tech corporations have found themselves in.

Lucas Maaser works as a Deputy Coordinator at the International Peace Bureau.

Stephanie Verlaan works for the International Peace Bureau as Deputy Coordinator and the Heidelberg Institute for International Conflict Research in an honorary capacity as Regional Director for Southeast Asia.

The US military is well-known as a world leader in developing advanced technologies. Indeed, it is widely credited with developing the Advanced Research Projects Agency Network or ARPANET, commonly thought of as the prototype of the internet. Other technologies initially developed for and by the US military, such as GPS or satellite imagery, have become revolutionary to how the average person goes about their daily life.

However, in the last decade, the US military has been toppled from its pedestal of technological innovation by the private tech industry’s major players. In order to remain on an equal footing with its adversaries, the Pentagon faced little alternative but to build a relationship with Silicon Valley of its own.

To maintain this relationship, government agencies have ample financial resources to provide. In 2020, the budget nations across the globe spent on their militaries reached 1,981 billion US dollars. With a 2.6-percent increase over the previous year, this development not only continues a long-standing trend of growing military budgets worldwide in the face of an ongoing pandemic, it further solidifies the US’s position as the unequalled front-runner on the list, with an estimated 778 billion dollars and 39 percent of global military spending.

Endeavours like Project Maven and the Joint Enterprise Defense Infrastructure (JEDI) illustrate both the US military's increased focus on artificial intelligence in the development of new, decision-centric military strategies, as well as the affinity of major tech companies to contribute to this effort. To do so, self-imposed ethical obligations like Google’s “Do No Evil” are cleverly circumvented for a share of the lucrative government contracts.

Between 2004 and 2021, the Department of Homeland Security (DHS) and the Department of Defense (DOD) alone invested more than 44 billion dollars in the services of Google, Amazon, Facebook, Microsoft, and Twitter. Silicon Valley’s ambitions to build close relations with government contractors are not limited by state borders, however, as their contribution to the European cloud computing initiative GAIA-X spearheaded by France and Germany demonstrates.

In this article, we set out to illustrate how some of the most lucrative contracts between big tech firms and the US military contribute to the building of war machinery, why we should care about this in Germany and Europe, and how people have organized in the past to make their voices heard.

Finding a Needle in a Haystack

Currently, one of the most prominent examples of Silicon Valley giants cooperating with the US military is Project Maven. The endeavour aims to autonomously identify “objects of interests” through the automated analysis of massive picture- and video data pools. Valued at an estimated 250 million dollars per year, it gained notoriety in 2018 when Google employees organized a walk-out to protest its employer’s support of the initiative, claiming a breach of the company’s ethical commitment to avoid actively contributing to technology with potential war applications.

In April 2017, then-Deputy Defense Secretary Bob Work started assembling an Algorithmic Warfare Cross-Functional Team. Its self-declared aim was to integrate artificial intelligence and machine learning technologies into existing DOD resources more efficiently to “maintain [its] advantages over increasingly capable adversaries and competitors” at the time. This endeavour could only be realized “with commercial partners alongside us”, said Marine Corps Colonel Drew Cukor, chief of the newly founded team at a 2017 presentation.

One of these commercial partners was Google, supporting the initiative by providing the Pentagon with their open-source AI software TensorFlow. “The US military does not use [TensorFlow] in weapons systems, and certainly not in supposedly autonomous ones. But the mere fact that Google is working with the US military has led to employee protests and ultimately to the company not renewing its existing contract with the Department of Defense”, said US security expert and senior fellow at the Center for a New American Security Paul Scharre in 2018.

As further damage control, Google CEO Sundar Pichai promised to “not design or deploy AI” in the area of “weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people” in a 2018 blog post titled “AI at Google: our principles”. As a reaction to Google cutting ties with Project Maven while continuing work on a search engine in line with the Chinese government’s policies, PayPal co-founder and Trump donor Peter Thiel deemed Google’s behaviour “seemingly treasonous”, a critique mirrored by Chairman of the Joint Chiefs of Staff Marine Corps General Joseph Dunford.

A recent analysis by Forbes associate editor Thomas Brewster revealed that Amazon Web Services Inc. (AWS) and the Microsoft Corporation filled the gap left by Google since 2019, securing a combined 50 million dollars in Pentagon contracts. He further found that despite his enhanced ethical commitment, Pichai becoming Google parent Alphabet Inc.’s CEO in late 2019 did not impact its thriving investments in start-ups like Orbital, Planet, and ClarifAI, which offer satellite image data and analysis services to the military through Alphabet’s venture capital wing GV. Not only is their field of expertise reminiscent of Google’s Project Maven contribution — Orbital evidently even won a contract under the Maven-umbrella, scoring 1.8 million dollars for the development of “high altitude still imagery multispectral models”.

With Rebellion Defense, Orbital is joined by another start-up working on Maven with direct links to the Mountain View-based tech giant. Founded by Innovation Endeavors, owned by Google’s former CEO Eric Schmidt, Rebellion Defense commits to “design AI products purpose-built for defence in the era in which software superiority will determine national security advantage”. Schmidt has built a reputation in bridging the gap between Silicon Valley and the defence sector, raising concerns over conflicts of interests through his membership in two DOD advisory boards for the enhancement of its AI technologies while retaining his role as a technical advisor at Alphabet and holding 5.3 billion dollars in shares of the Google parent.

Rebellion Defense CEO and co-founder Chris Lynch also holds a role as founding director at Defense Digital Service (DDS), a Pentagon “rapid response team” that initiated JEDI — an initiative best-known for the legal stalemate it created between Google and AWS over the 10-billion-dollar project. Lynch’s founding partners Nicole Camarillo and Oliver Lewis both share his extensive background in the security sector.

This dynamic not only feeds into a history of Google actively following the money to Pentagon contracts while working hard to maintain their image as a “Do No Evil” consumer tech innovator — it is also indicative of the growing revolving-door hiring practices between the US defence sector and Silicon Valley.

The Force in the Cloud

Even before attracting attention over its legal dispute, the Joint Enterprise Defense Infrastructure (JEDI) project signalled another major area of development for which the DOD sought assistance from major US tech firms. JEDI was a multi-cloud computing project that aimed to support enhanced communication between the Pentagon and soldiers in the field as well as between the different defence agencies. The contract, for which AWS was the anticipated awardee, was ultimately assigned to Microsoft, which AWS later contested in court.

Then in July 2021, the DOD announced it was dissolving the 10-billion-dollar JEDI contract, reasoning that it no longer met the Department's needs. In reality, the basis of this decision was a reflection of how rapidly the DOD’s needs change and the legal battle between the vendors had pushed the deliverable timeline beyond an acceptable limit.

Immediately after cancelling JEDI, the DOD announced the Joint Warfighting Cloud Capability (JWCC) project as its replacement. The JWCC differs from JEDI in that rather than developing a single cloud entity via a single vendor, the contract would be awarded to multiple Cloud Service Providers that DOD deemed capable of meeting its requirements. Each vendor would be tasked with developing a cloud entity with a specific purpose. This move intended for each cloud to be better readily protected and make containing security breaches easier.

Awarded contracts would take the form of “indefinite delivery indefinite quantity” (IDIQ), stipulating an indefinite amount of services within a specific timeframe. The DOD announced in November 2021 that it expects to award IDIQ contracts each to Microsoft and AWS. It has also put out a request for proposals to Google Cloud and Oracle to join the initiative.

It is here that the controversy begins. The walkout staged by Google employees over its involvement with Project Maven set the precedent for the company’s decision not to bid for the JEDI contract. Google Cloud CEO Thomas Kurian explained in a blog post dated 12 November 2021 that, although its decision not to be involved with JEDI was due to a lack of assurance its work would not violate its AI ethics principles, he believed joining the JWCC would not violate this code. He did acknowledge, however, that it was understood that not all employees would agree with this reasoning.

Kurian attempted to justify this by differentiating between JEDI and JWCC and highlighting other defence-related projects it had successfully collaborated with DOD on without violating its ethics code. Kurian’s blog post was a response to questions raised in a company-wide staff meeting, during which over 1,000 employees questioned Google's involvement in JWCC. At the time of writing, AWS had not reported any internal controversy or ethical concerns surrounding its involvement with either JEDI or JWCC.

Big Companies, Small Players

Of the major tech companies the DOD is seeking to collaborate with, Apple appears to be a relatively small player.

In 2015, it became a member of a consortium of 162 companies called the FlexTech Alliance, which was commissioned to develop various hardware technologies for defence purposes. The 75-million-dollar project aimed to develop flexible electronic systems that could be embedded in materials such as silicon and were lightweight enough to be worn by soldiers, yet resilient enough to be moulded onto the outside of aircrafts.

While Apple had not partnered with the military prior to joining the Alliance, its senior management was concerned that the company’s profit growth margin would begin to stagnate if it did not explore new and alternative markets to expand into. Traditionally, the sale of its iPhone represented the vast majority of its profits. With the market becoming increasingly flooded as its competitors rose to meet its bar, however, it needed to break out into alternative product markets.

In a similar vein, with a total of 365,000 dollars in DHS contracts and 170,000 dollars in minor DOD contracts, Meta Platforms Inc. — previously known as Facebook — joins Apple as a relatively small player in the business of military contracting.

Helping Europe Grow into Data Sovereignty

Shifting the view from the US to Europe, it is apparent that the expertise of Silicon Valley tech firms is equally crucial to realize projects of a certain scale outside American borders, as the project GAIA-X exemplifies.

Initiated in 2019 “with the goal of developing a trustworthy and sovereign digital infrastructure based on European rules”, GAIA-X is the German-French attempt to build a European open-source alternative to the cloud computing technologies of US and Asian tech giants that dominate the market. To realize the endeavour, 22 partners from the tech, economic, and science sector initially founded the Belgian non-profit organization GAIA-X European Association for Data and Cloud AISBL in September 2020, bringing together major German players like BMW, Bosch, Deutsche Telekom, the Fraunhofer Gesellschaft, SAP, and Siemens.

While military applications are not communicated as a major purpose of the project, the IT provider of the German Army, BWI, describes GAIA-X as a capable resource for armed forces to maintain “the necessary control and action options in the cyber and information space in order to be able to fulfil their constitutional mission — self-determined and free from unwanted third-party influence.”

Since its inception, the number of parties to the initiative rapidly increased to over 300, attracting criticism for including Chinese members like Huawei or Alibaba as well as major US firms. “While there is a lot of talk about cloud sovereignty, current plans by governments in Europe as part of GAIA-X still rely on US technologies by AWS, Google and Microsoft which are subject to foreign surveillance”, comments the European Cloud Industrial Alliance (EUCLIDIA) founded by 23 companies in 2020 monitor development.

Even bigger concerns were raised by the involvement of big data integration corporation and US spy agency contractor Palantir, which was announced to have joined “GAIA-X as a proud Day 1 Member” in December 2020, a move observers said “should at least raise an eyebrow for people in Europe.”
Considering the stark voices of protest by a political and corporate opposition, one might find that it did — one of them being Anne Roth, an internet policy consultant for Die Linke, who commented on Twitter: “There goes the trust in European sovereignty.”

Co-founded by Peter Thiel, Palantir has built a reputation developing AI tools to aid both autocratic and liberal democratic governments in the surveillance of their citizens and borders. Christopher Soghoian, a technician at the American Civil Liberties Union, dubbed it “a key force in the surveillance-industrial complex”. GAIA-X is not the surveillance corporation’s first major contract in Europe. Since 2016, Europol has used their anti-terror tool “Gotham” for their investigations. The software has also been used by the German federal state of Hesse under the name “Hessendata” since 2017.

“To some degree, I think there’s a focus on an almost colonial behaviour from these American companies”, comments Jack Poulson, former Google data scientist and founder of Tech Inquiry, a non-profit focusing on the analysis of contract procurement streams to render ties between private tech corporations and the US government more transparent. “Obviously there are human rights issues everywhere — whether that’s Palantir coming in and wanting to sell technology to amplify border surveillance or deportations. It’s certain these companies have close relationships with the intelligence communities across Europe.” Concerning US tech firms’ investments in enhancing these relationships, Poulson elaborates: “I don’t really see why it would develop fundamentally differently than in the US, even if one might expect that Europe is going to try and create its own analogues of these companies.”

Even before Palantir’s GAIA-X announcement, Sophie in ‘t Veld, Dutch Member of the European Parliament and contributor to the Renew Europe Group, stated in October 2020: “A company with Palantir’s track record should not be considered as a partner for any EU-wide project, and the European Commission knows it. This secretive corporation is at odds with the European values many EU-citizens hold dear, such as privacy, civil liberties, and transparency of government — not to mention the strategic implications of cooperating with an American intelligence contractor.”

Besides data sovereignty concerns, recent reports show the project suffering from over-bureaucratization, a lack of focus, and confusion over conflicting interests, resulting in French cloud provider Scaleway not renewing its contract with the project in November 2021.

Organizing the Opposition

The pool of evidence indicating a strong connection between major US tech firms and military projects both nationally and globally raises the question of how to sustainably address its consequences. While there may be a wide variety of potential paths to do so, following the journey of one former tech employee aiding the analysis and publication of potentially harmful contracts might best highlight the struggles and pitfalls connected to such an endeavour and resources needed to work past them.

Virtually all available investigative reports on the subject matter are based on information provided by Poulson’s non-profit Tech Inquiry. After his 2018 resignation over Google’s (later cancelled) censored search engine project in China, the transition to the non-profit sector was not immediate, he recalls: “For about the first year-and-a-half, we were about as Luddite as you could be. Our entire focus laid on speaking with journalists, raising awareness over issues that folks in the non-profit sector had been familiar with.”

The approach to monitor both conditions and connections between contractors developed out of an observation in the industry: “When I met with some of the more senior officials, I found that human rights issues were not much of a concern of theirs. It was much more about maintaining a close military relationship with the US tech companies”, he recalls. “As part of that, I started to become more aware of just how much bureaucracy there was surrounding those relationships, whether it’s the Defense Innovation Board, the Defense Innovation Unit, the National Security Commission on AI, In-Q-Tel etc. I began to do Freedom of Information requests into some of those relationships, and came to the conclusion that there are a lot of relationships between these companies that are just not very well documented.”

Tech Inquiry emerged out of this practice, harbouring other former employees from the tech industry both as board members and contributors to the non-profit’s research. While their work does not explicitly focus on disarmament, some members bridge the gap to pacifist movements through their engagement in anti-militarist campaigns. One prominent example is the Campaign to Stop Killer Robots, a major global initiative calling for a ban on lethal autonomous weapons. “Several of the members and board members of Tech Inquiry have played a major role in it. For example, Liz O'Sullivan had left ClarifAI over their work on drone surveillance. Lauren Nolan had left Google for the same reason over their work on Project Maven.”

Finding Answers in Unexpected Places

Limiting connections between major US tech companies and the US government to their impact on military actions, however, is not particularly helpful to grasp the reality of how tech companies conduct their business, Poulson suggests: “I tend to find that if you’re looking from the perspective of tech companies, they’re really just trying to sell their technologies everywhere. There’s another lens where one can view a lot of military technology as being upstream from technology that will end up in the hands of the Department of Homeland Security through drone surveillance, autonomous border surveillance, facial recognition, etc. And then it works its way into more municipal usage.”

Widening the scope of analysis to a broader range of contracts may reveal issues in areas that might otherwise be left unconsidered, Poulson finds: “I think if you do a lot closer monitoring of procurement and have a deeper understanding of the bureaucracy of the US government, you understand that the intelligence community works very closely with a lot of agencies people don’t tend to think of.”

This practice further helps re-contextualize the purported ethical neutrality of contracts within a public discourse. One area this may have a particular impact on is cloud computing: “One of our big wins was showing that Google Cloud was being sold to a company called Thundercat Technology, which is a fairly prolific contractor with US Customs and Border Protection. We found evidence that Thundercat planned to use Google Cloud’s AI to process thermal imagery from Anduril Industries’s autonomous border surveillance, which pretty well contradicts the way Google has positioned itself in terms of what it would and would not do at the US border. This raises an interesting question considering the role cloud computing plays financially and technologically. It’s both a huge component of defence sales and a driving source of revenue for major tech companies.

If these tech companies frame cloud computing as morally neutral, which is what we saw post-Maven, then how do we actually show what they’re actively contributing? What levers for accountability really exist at that phase besides, say, employees leaking internal communications? I think the more we can show what these so-called ‘black box’ cloud contracts actually result in, the more we can push back on the narrative that this material support for militaries and intelligence agencies around the world is not just the equivalent of selling them a pile of steel, that there’s an actual direct contribution to their activities.”

Erecting Structures for a Sustainable Counter Movement

As an employee, the decision to object to practices perceived as unethical often creates a field of tension between doing what’s right and pursuing a long-term career — a conflict that Tech Inquiry attempts to alleviate through the protection of whistle-blowers’ identities. Even without this layer of security, there are opportunities for tech workers to speak out, however: “If nothing else, you don’t have to put your name on something. If you want to share something with the press, you can do so anonymously”, Poulson explains.

Even with additional resources provided by the non-profit sector, the pressure emerging from this individual conflict may prompt employees to opt for a less confrontational approach towards addressing issues they see — a practice that is often used as an opportunity for executives to stay inactive, Poulson finds. “This also gets to the heart of a debate over organizing versus speaking out in protest. I think organizing is obviously one of the more sustainable paths. It also, in a sense, builds power. I think it should come up more in circles relating to whistleblowing. And it’s something I’ve pushed a lot for.

I think one of the pitfalls I experienced is that a lot of the senior people I knew weren’t doing anything close to what we would traditionally think of as organizing. The ‘change from the inside’ narrative is often used as an excuse for why it is okay for them to do nothing. I feel like that side of the conversation isn’t talked about much. One question might be how to prevent people from getting away with the bad faith application of a ‘change from the inside’ narrative”, a process in which unionizing can play a crucial role.

According to Poulson, however, unions can only serve as one contributor to a much more holistic approach towards addressing these issues: “The complication is that unions often represent the interests of the workers, which can dramatically differ from the interests of the public. So I think it’s always worth emphasizing that beside unions, we also call for coalitions, including independent civil society organizations.” This also means a better organization of collaborative efforts between non-profits in the sector, he suggests. ”That doesn’t mean not ever criticizing. But I think when non-profits can come together and work on projects that combine their strengths instead of just competing with each other, it tends to lead to incredible impact.”

To do so sustainably, however, support from the wider public is needed: “I think one of the major problems in the non-profit space is that most of the money comes from the very tech billionaires that you’re trying to critique. If we genuinely want to tackle tech billionaire influence, we really need to be able to be self-sustaining.

Even at the most respected organizations, you constantly see even high-level figures moving into tech companies because there’s so much power and honestly, non-profits just can't usually pay very well. So I think there’s no real substitute for just making civil society a place in which people can actually have a career, and that’s going to ultimately come from taxes and before that from grassroots donations.”

In order to acquire these means and forge meaningful alliances to build a sustainable and organized civil society corrective to the tech sector’s business of war in Germany and Europe, we first need to create spaces of discourse and exchange. These are spaces that define war both interdisciplinarily and intersectionally, that allow for the consideration of perspectives of those affected most by the technology of war — from communities in active war zones over refugees dying at the European borders to groups particularly marginalized and policed domestically.

Only out of this discourse can a movement can emerge that holds tech companies accountable for their actions on the basis of human rights, not self-imposed ethical obligations. This is a movement that does not limit itself to targeting Google for circumventing its ethical guidelines, but raises the question of why more companies are not formulating them in the first place.

Technology is a human invention — its purpose and use is controlled by us. With this in mind, there is no basis to claim that it is neutral. Technology has always been political and will continue to be. In a world where every person using consumer-targeted services engineered by Silicon Valley corporations adds to their data pool and revenue, there is no excuse to remain inactive — in the US, in Europe, and around the world.