The connections between big tech and militaries 

On the morning of 28 February 2026, several American missiles hit a primary school in Minab, during the first day of the surprise attack on Iran by the US and Israel. The strikes on the school killed 168 people, most of them children.

Almost immediately, people asked whether Claude, an artificial intelligence (AI) tool owned by Anthropic, had marked the school as a target. Claude was, until then, mostly known as a chatbot, but it was also the only large language model allowed inside the US’s classified military systems.

Security briefings revealed that the school had been targeted because a Maven Smart System database that Claude used for intelligence had been out of date. Controversial data analysis company Palantir, which runs Maven, had not uploaded the information that the Islamic Revolutionary Guard no longer inhabited the building, and that by 2016 at the latest it had been turned into a school, separated from their compound.

UK military and novel tech 

Taking into account inflation, the UK spends more on the military now than it did in the 1980s, at the height of the Cold War arms race. 

It budgeted £60bn for defence in financial year 2024/25. This will go up to £73.5bn in 2028/29 – £7bn more than the government plans to spend on schools – to grow the defence industry.

As part of this strategy, the Ministry of Defence will use at least 10% of its equipment budget to buy “novel technologies, such as quantum technologies, uncrewed and autonomous systems, and AI” – with the help of “SMEs, start-ups, dual-use tech companies, and private finance”.

Historical military tech links

Close relationships between tech companies and the military are not new.

ARPANET, the first version of what we now know as the internet, was largely funded by the US Department of Defense’s Advanced Research Projects Agency (DARPA). Defence and intelligence agencies also put up much of the funding for the early development of Silicon Valley as a home for tech.

It does not come as a surprise then that the Big Tech companies, already nestled into most aspects of our daily life, are also deeply intertwined with the world’s military systems.

Little tech and the military

The military first collaborated with companies making consumer tech products on basic programmes such as web browsers and email clients.

It is ironic that some of the open source and privacy-conscious products we now see as alternatives to Big Tech were created by or with the military in the 2000s. The French Ministry of Defence, for example, started using open source software such as Mozilla’s Firefox web browser early on. It also donated security features to Mozilla’s Thunderbird, to make the email client safe to use for the military.

The Tor browser, favoured by journalists and activists as well as Dark Web drug dealers and CIA agents, was developed by the US Naval Research Lab (NLR), to “create internet connections that don't reveal who is talking to whom, even to someone monitoring the network”.

Tor’s decentralised network of servers and encryption is designed to bury traffic that people want to keep secret in a pool of as many anonymous users as possible.

To this day, NLR researchers are part of the Tor team and the Tor Project’s latest annual report lists DARPA as one of its funders – although we did not find evidence that this collaboration went further than developing "unobservable and compromise-resistant obfuscation channels … for censorship-resistant communication”.

We write more about Tor in our web browsers guide.

Enter Big Data: Project Maven

In 2015, the Pentagon opened a Silicon Valley start-up accelerator, DIUx, to find new technologies that could be useful to the military. It was specifically looking to invest in companies working on AI, robotic systems, big data analysis, cybersecurity, and biotechnology.

The US Department of Defense (DoD) was frustrated that Big Tech giants were far ahead of it in their use of large datasets. As Google Maps directed drivers around traffic jams, Donald Trump and Brexit campaigners focused on Facebook profiles that would be most receptive to their messages, and Alexa encouraged people to shop online, the US military wanted to turn real-world big data into automated military action with Project Maven.

The first aim of Project Maven was to make sense of the overwhelming amount of video the department captured during its field operations. The DoD directed DIUx to find corporate partners that could develop the algorithms needed for this. Within months, it was using them for drone missions against ISIS in Iraq and Syria.

Google was the first Big Tech company involved in Project Maven. With AI, it analysed the drone footage and was working towards drone missions that did not require human direction. Google workers protested against the way the company was going. 4,000 employees signed a letter stating that “Google should not be in the business of war”, and a dozen workers resigned. Because of this, Google did not renew its Project Maven contract.

In 2026, Project Maven has become Maven Smart Systems, run by Palantir. It is directing the war in Iran as the core AI operating system of the US military, and NATO is using it to prepare for “the future of warfare”. 

Maven can track targets in real time, recommend which weapons to use, and has reduced the time between detection and shooting “from hours to minutes or less”.

The Palestine laboratory

Before Maven could fly this high, the Israeli military battle-tested similar systems in Gaza during the ongoing genocide. Reporters have uncovered three AI-powered tools: The Gospel, Lavender, and Where’s Daddy? An American magazine, The Nation, found that Palantir had likely provided the know-how needed to build these capabilities.

Israeli officials claim the country fought its “first AI war" in 2021, when it bombed Gaza for 11 days, likely already with help of The Gospel. The Gospel locates targets like underground tunnels, family homes of suspected militants, and civilian infrastructure that when destroyed is meant to shock people so much they “put pressure on Hamas”.

By the time Hamas attacked army outposts and surrounding villages in Israel on 7 October 2023, Lavender and Where’s Daddy? were also ready for deployment.

Israel had relied increasingly on surveillance technology instead of soldiers to confine Palestinians within Gaza’s land, sea, and air borders. The Al Jazeera news network found, in 2023, that analysts have even suggested that over-confidence in this technology contributed to intelligence agencies ignoring repeated warnings that Hamas was preparing a break-out.

In the early stages of Israel’s genocidal campaign, Lavender played a central role in the relentless bombing of Palestinians.

Its official task was to find suspected militant members of Hamas and Islamic Jihad, including low-ranking ones, and mark them as potential bombing targets.

In practice, the military used it as an almost instant kill-list, only checking if the target was male before clicking ok.

Where’s Daddy?, a system as cynical as its name, tracked people on the Lavender list and let military know when they entered their family home. Intelligence sources interviewed by +972 Magazine stated this made it easier to locate them rather than during military activity. As a result, they were usually bombed inside residential blocks, together with their family – killing thousands of Palestinians who had nothing to do with military activity.

Israeli weapons manufacturers, state-owned and private, have a well-documented history of selling to other countries the weapons and surveillance technologies they use to control and kill Palestinians. Author Antony Loewenstein calls this “the Palestine laboratory”: Palestinians are the test subjects that allow Israel to hone its military tools that it then sells as “battle-tested”. 

As these tools become more digital, the companies that make the tools run become part of this supply chain.

Tech for genocide

It is clear that running these systems requires significant digital infrastructure. Several companies are known to have large-scale contracts with the Israeli government, but because AI and cloud services can be used for any type of work it can be hard to prove that they are directly complicit in genocide.

An investigation by AP news agency found that after 7 October 2023, the Israeli military’s use of Microsoft and OpenAI technology “skyrocketed”.

Advanced AI models provided by OpenAI used Microsoft’s Azure cloud services to aggregate information gathered from mass surveillance of Palestinians in Gaza. The data was then matched up with Israel’s “in-house targeting systems”.

The Guardian, +972 Magazine, and Local Call, who published groundbreaking investigations into Israel’s use of tech to perpetrate genocide, reported this to Microsoft. As a result of the scrutiny, Microsoft announced that it had ended certain Israeli military subscriptions and services, without going into detail on who could still use its systems for what.

An investigation by Wired magazine found that there was also a much closer link between the Israeli military and Google and Amazon than their official communications lead to believe.

Google and Amazon workers had protested against Project Nimbus since their employers signed a contract to
provide large-scale cloud services to the Israeli government in 2021, which they suspected were made available for military use. Wired found documents and statements that confirmed the Israeli military had been “central to Project Nimbus since its inception” and top Israeli officials appeared to think it provided “important infrastructure for the country’s military”. According to the article, Google continued to insist that Project Nimbus was not “relevant to weapons or intelligence services,” while Amazon seemed to not publicly discuss the scope of the contract.

More on Amazon’s involvement in the infrastructure of genocide can be found in our dedicated web article.

Can we challenge big tech's involvement with the military?

It doesn’t look like Big Tech will stop trying to cosy up to the military any time soon. 

Just days before the mistaken attack on the Iranian school, Anthropic blocked the Pentagon from using its Claude AI for autonomous weapons and mass surveillance in the US. But hope that this would change the course for military use of AI was dashed when SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, and Amazon Web Services jumped in and made their technology available to the Pentagon for “any lawful use”. 

Big military contracts are very profitable, and AI companies are already struggling to pay for the data centres and training needed to stay ahead in the race for dominance of the AI market.

But there are glimmers of hope. Below are different campaigns and resistance activities from tech workers, regular consumers and civil society.

(a) Tech workers

As Big Tech is integrating into real-time war systems, regular tech workers at Google or Microsoft become responsible for maintaining them. They might not care much if an out-of-date database loses a client a week’s work, but likely do care if it kills over 160 people, mostly children.

Not surprisingly then, amongst the most vocal campaigners against Big Tech’s involvement in war are its employees. 

No Tech For Apartheid and No Azure For Apartheid are organising their co-workers at Google, Amazon, and Microsoft to refuse being made complicit. 

In the UK, 98% of workers at Google DeepMind, an AI frontier lab, voted to unionise with the United Tech & Allied Workers Union, “to end use of their technology by Israel and the US military”. They have already threatened “research strikes”.

(b) Civil society

The Challenging Corporate Power feature mentions the 300+ groups campaigning for change around military uses of AI. Many of these are under the umbrella of the stopkillerrobots.org alliance. They are calling for a new international law on autonomous weapons systems. This new treaty should prohibit autonomous weapons systems which target humans directly, and autonomous weapons systems which cannot be used with meaningful human control.

(c) Consumers and purchasers

Ethical Consumer has devised a new 'military links' ranting for the guides in this issue so that consumers and institutions such as universities and public bodies can take into account the degree of involvement before deciding to use these tools. 

This military rating assesses whether tech companies had any direct or indirect links with the military, and rewards companies which had explicit policies prohibiting military contracting.

More information about how each of the tech companies score on this issue appears in the guides:

• social media
• search engines
• web browsers

The AI tools guide goes into much more detail on this issue, and discusses the issues of state surveillance and Palantir too.

And, if all this fails, Big Tech may learn the hard way that joining the military makes it a military target. Iran has already bombed, deliberately, three Amazon data centres in the United Arab Emirates and Bahrain. Two months later, retailers and logistics firms were still waiting for their Amazon Web Services to go back to normal.

About the header image

The image is by Kathryn Conrad and is available from Better Images of AI, a non-profit collaboration, who are 'researching, creating, curating and providing Better Images of AI', rather than cliched and misleading stereotyped images.