The standoff between Anthropic and the Pentagon has forced the tech industry to once again grapple with the question of how its products are used for war – and what lines it will not cross. Amid Silicon Valley’s rightward shift under Donald Trump and the signing of lucrative defense contracts, big tech’s answer is looking very different than it did even less than a decade ago.
Anthropic’s feud with the Trump administration escalated three days ago as the AI firm sued the Department of Defense, claiming that the government’s decision to blacklist it from government work violated its first amendment rights. The company and the Pentagon have been locked in a months-long standoff, with Anthropic attempting to prohibit its AI model from being used for domestic mass surveillance or fully autonomous lethal weapons.
Anthropic has argued that giving in to the DoD’s demands to permit “any lawful use” of its technology would violate its founding safety principles and open up its technology for potential abuse, staking an ethical boundary that others in the industry must decide whether they want to cross.
Although Anthropic’s refusal to remove safety guardrails and the Pentagon’s subsequent retaliation have highlighted longstanding concerns over the use of AI for conflict, the fight has shown how much the goal posts have moved when it comes to big tech’s ties to the military.
“If people are looking for good guys and bad guys, where a good guy is someone who doesn’t support war,” said Margaret Mitchell, an AI researcher and chief ethics scientist at the tech firm Hugging Face. “Then they’re not going to find that here.”
Anti-military protests to military contracts
There’s a number of contributing factors in big tech’s newfound embrace of militarism. Its alignment with the Trump administration, which has included shows of fealty to Trump from major CEOs, has tied tech firms to the government’s desire to expand its military capabilities. The administration’s vow to overhaul federal agencies using artificial intelligence has also specifically signaled an opportunity for AI firms to integrate their products into government and military operations in a way that could secure revenue for years to come. Looming in the background, concern over China’s technological advancement and a surge in international defense spending have also shifted attitudes in the industry.
It was not so long ago, however, that working with the military on potentially harmful technology was seen as a red line for many big tech workers. In 2018, thousands of Google employees launched a protest against a program to analyze drone footage for the DoD called Project Maven.
“We believe that Google should not be in the business of war,” over 3,000 workers stated in an open letter at the time. Google decided not to renew Project Maven following the protests and published policies that barred pursuing technology that could “cause or directly facilitate injury to people”.
In the years since the Project Maven protest, though, Google has clamped down on employee activism, removed the 2018 language from its policies that prohibited creating technology for weaponry and signed numerous contracts that allow militaries to use its products. In 2024, the tech giant fired over 50 employees in response to protests against the company’s military ties to the Israeli government. Chief executive Sundar Pichai sent a memo to employees after the firings stating that Google was a business and not a place to “fight over disruptive issues or debate politics”.
Google announced just this week that it would provide its Gemini artificial intelligence to provide the military a platform for creating AI agents to work on unclassified projects.
OpenAI too had a blanket ban on allowing any militaries to access its models prior to 2024, but since and now has its chief product officer serving as a lieutenant colonel in the US military’s “executive innovation corps”. The startup, along with Google, Anthropic and xAI, signed an up-to-$200m contract with the DoD last year to integrate its technology into military systems. On the day that Pete Hegseth, the defense secretary, declared Anthropic a supply chain risk, OpenAI secured a deal with the DoD allowing its tech to be used in classified military systems.
Elsewhere in the tech industry, more hawkish companies like defense tech firm Anduril, founded the year before the Google Maven protests, and surveillance tech maker Palantir have made partnering with the DoD a cornerstone of their businesses and attempted to sway Silicon Valley politics towards their worldview. Palantir has been ahead of the curve on working with the military, contracting with military intelligence to map planted explosives in Afghanistan in the early 2010s. Chief executive Alex Karp published a book last year dedicated in large part to advocating for closer integration of the tech industry and AI with the US military, in one passage accusing the Google workers who protested in 2018 of being nihilists.
After Google dropped the Project Maven contract in 2019, Palantir took it over. Maven is now the name of the classified system that military personnel use to access Anthropic’s Claude, according to the Washington Post.
Anthropic goes to war
Even as Anthropic has received public praise in its standoff with the Pentagon, its co-founder and chief exceutive Dario Amodei has emphasized that the AI company and the government largely want the same things.
“Anthropic has much more in common with the Department of War than we have differences,” Amodei wrote in a blogpost last Thursday.
While the White House has accused Anthropic of being “a radical left, woke company”, Amodei’s views on the use of AI in conflict and fears of its misuse are far from tree-hugging pacifism. In a lengthy essay published in January, Amodei warned against potential harms of AI such as the creation of deadly bioweapons and threats from China maliciously using the technology. Simultaneously, he argued that companies should arm democratic governments and militaries with the most advanced AI possible to combat autocratic adversaries.
He expressed less concern about AI making it easier to kill people or conduct warfare and more about the reliability of the technology and threat of it being consolidated by too small a number of people with “fingers on the button” who could control an autonomous drone army.
Amodei’s essay also foreshadowed some of the central issues involved in his fight with the Pentagon, including the potential for AI as a tool of mass surveillance. While arguing for bulwarks against the abuse of AI, he stated that his formulation was that it was okay to use the technology for national defense “in all ways except those which would make us more like our autocratic adversaries”.
While Amodei has so far stuck to the company’s red lines, he has also repeatedly stated that he wants Anthropic to continue working with the Defense Department. The company’s lawsuit against the DoD showcases how extensively the company has been willing to work with the military and alter its products for their use.
“Anthropic does not impose the same restrictions on the military’s use of Claude as it does on civilian customers,” Anthropic’s California lawsuit stated. “Claude Gov is less prone to refuse requests that would be prohibited in the civilian context, such as using Claude for handling classified documents, military operations, or threat analysis.”
The government has reportedly been using Claude for target selection and analysis in its bombing campaign against Iran, a use-case that Anthropic has given no indication that it has an issue with. In his blog post on Anthropic’s website last week, Amodei stated that he did not believe that his company had any role in the military’s operational decision-making. He claimed that Anthropic supports American frontline warfighters and remains committed to providing them with technology.
“We have said to the department of war that we are OK with all use cases,” Amodei told CBS News last week. “Basically 98 or 99% of the use cases they want to do, except for two.”
Facts Only
Anthropic, an AI firm, sued the U.S. Department of Defense (DoD) three days ago, claiming its blacklisting from government work violated its First Amendment rights.
The dispute centers on Anthropic’s refusal to allow its AI model to be used for domestic mass surveillance or fully autonomous lethal weapons.
The DoD demanded Anthropic permit “any lawful use” of its technology, which the company argues violates its safety principles.
In 2018, over 3,000 Google employees protested Project Maven, a DoD program analyzing drone footage, leading Google to drop the contract and adopt policies against weaponized AI.
Google has since reversed its stance, removing language prohibiting weaponry-related technology and signing military contracts, including a recent deal to provide its Gemini AI for unclassified military projects.
Google fired over 50 employees in 2024 for protesting its military ties to Israel.
OpenAI previously banned military use of its models but now allows it, with its chief product officer serving as a lieutenant colonel in the U.S. military’s “executive innovation corps.”
Anthropic, Google, OpenAI, and xAI signed a contract with the DoD in 2023 worth up to $200 million to integrate AI into military systems.
Palantir, a surveillance tech firm, has long worked with the military and criticized Google’s 2018 protests as “nihilistic.”
Anthropic’s CEO, Dario Amodei, stated the company shares most goals with the DoD but opposes domestic mass surveillance and autonomous lethal weapons.
The U.S. military uses Anthropic’s Claude AI for target selection in bombing campaigns against Iran, a use case the company has not publicly opposed.
Anthropic’s lawsuit reveals it has modified its AI for military use, making it less restrictive than its civilian version.
The Trump administration has pushed for AI integration in federal agencies, creating opportunities for tech firms to secure defense contracts.
Executive Summary
The standoff between Anthropic and the Pentagon highlights a shifting landscape in Silicon Valley’s relationship with the military. Anthropic’s lawsuit against the Department of Defense (DoD) stems from its refusal to allow its AI technology to be used for domestic mass surveillance or autonomous lethal weapons, a position it argues aligns with its ethical principles. This dispute contrasts with the broader tech industry’s growing willingness to engage with military contracts, a trend accelerated by the Trump administration’s push for AI integration in defense operations.
Historically, tech workers resisted military collaboration, as seen in Google’s 2018 employee protests against Project Maven, which led the company to drop the contract and adopt policies against weaponized AI. However, Google has since reversed course, signing military deals and suppressing employee dissent. Similarly, OpenAI, once opposed to military use of its models, now permits it and has deepened ties with the DoD. Companies like Palantir and Anduril have long embraced military partnerships, framing them as necessary for national security. Anthropic’s partial resistance—opposing some uses while cooperating on others—reflects the industry’s complex balancing act between ethical concerns and lucrative defense contracts.
Full Take
The strongest version of this narrative frames the tech industry’s evolving stance on military collaboration as a pragmatic response to geopolitical pressures and economic incentives. The article credibly documents the shift from resistance (e.g., Google’s 2018 protests) to acquiescence, driven by factors like China’s technological rise, government demands, and corporate suppression of dissent. It also highlights Anthropic’s nuanced position—opposing specific uses while still engaging with the military—as a rare attempt to draw ethical boundaries in an industry increasingly aligned with defense interests.
However, the narrative risks oversimplifying the moral stakes. The framing of tech firms as either "good guys" (ethical resistors) or "bad guys" (military collaborators) obscures the reality that most companies, including Anthropic, are negotiating compromises rather than taking absolute stands. The article also leans into a false binary—suggesting that opposition to military AI is either pacifist idealism or nihilism (as Palantir’s CEO claims)—without exploring middle-ground perspectives, such as conditional cooperation with safeguards.
Root cause: The paradigm driving this narrative is the tension between profit motives and ethical constraints in an era of escalating U.S.-China competition. The unstated assumption is that military AI is inevitable, leaving tech firms to choose between resistance (with potential reputational or financial costs) or compliance (with moral trade-offs). This echoes historical patterns of corporate-military convergence during Cold War-era tech development, where ethical concerns were often sidelined for strategic advantage.
Implications: The erosion of tech workers’ ability to challenge military contracts—exemplified by Google’s firings—undermines internal accountability. Meanwhile, the normalization of AI in warfare (e.g., target selection) raises questions about human agency: Who bears responsibility when algorithms assist in lethal decisions? The beneficiaries are clear—defense contractors, tech firms securing lucrative deals, and governments expanding surveillance capabilities—but the costs (e.g., civilian casualties, erosion of democratic oversight) are diffuse and long-term.
Bridge questions: What would it take for tech workers to regain leverage in shaping corporate ethics? How might international treaties or industry-wide standards mitigate the risks of autonomous weapons without stifling innovation? If Anthropic’s "red lines" are breached, what precedents does that set for other firms?
Counterstrike scan: A coordinated influence campaign would likely amplify the "China threat" narrative to justify military AI, while dismissing ethical concerns as naive or unpatriotic. The article does reference geopolitical competition but avoids overt fear-mongering, focusing instead on documented shifts in corporate behavior. No structural alignment with a hypothetical attack playbook is detected.
Patterns detected: ARC-0024 Ambiguity (false binary of "good guys vs. bad guys"), ARC-0043 Motte-and-Bailey (Anthropic’s selective opposition to military uses while still cooperating).
Sentinel — Human
The article exhibits strong human signals, including stylistic diversity, specific attributions, and nuanced arguments, with no significant indicators of synthetic origin.