Both the Trump Administration and Big Tech are Bound by Obligations to Protect Civilian Life

CONTACT: 
Catalina Brennan-Gatica | catalina@theworkeragency.com | 203-654-0071

San Francisco, CA— Yesterday, Tech Justice Law, alongside an international coalition of other human rights and technology justice organizations – Abolitionist Law Center, Access Now,  Center for Constitutional Rights, and the Centre for Research on Multinational Corporations (SOMO) – filed an amicus brief in support of neither party in Anthropic PBC v. U.S. Department of War, et al., where Anthropic’s motion for summary judgment is currently pending before the U.S. District Court for the Northern District of California (San Francisco Division).

“Anthropic and the Department of War act as though emerging technologies are beyond the reach of the law. They are not,” said Maddy Batt, Legal Fellow at Tech Justice Law. “The U.S. must meaningfully distinguish between civilians and combatants when attacking, and outsourcing that assessment to an unreliable technology in order to kill within seconds leads to war crimes.”

As the Court deliberates on Anthropic’s motion, this brief seeks to refocus the dispute on the legal obligations both the U.S. government and the company and its corporate offices have under U.S. criminal and international humanitarian law. Anthropic and the Trump administration’s underlying disagreement centers on whether Anthropic’s AI-program Claude should be deployed fully autonomously, largely disregarding the existing legal constraints governing the use of militarized AI. In light of the existing use of semi-autonomous AI in armed conflict, including in Iran, in ways that appear to violate the law of armed conflict, Amici argue to the Court that the judiciary should explicitly remind both parties of their obligations under international law.

“Both parties are bound by U.S. and international humanitarian law, and technological advancements compressing the ‘kill chain’ – including and especially during armed conflict – demand the Court’s careful, intensive review to ensure adherence to the most fundamental principles of law,” said Sadaf Doost, Staff Attorney and Human Rights Program Manager at Abolitionist Law Center.

The brief explains that, even when deployed without full autonomy, AI systems synthesize terabytes of data–one terabyte of which is equivalent “to the contents of about one million books”– to produce recommendations for attacks that human operators cannot meaningfully review. The speed and scale with which strikes are executed in Iran demonstrates that militaries are de facto deferring to AI outputs, without meaningfully assessing the legality of the attack, in violation of international humanitarian law and the U.S. War Crimes Act.

“Anthropic, its corporate officers and the Department of Defense/War are not constrained merely by their own views on militarized AI and have demonstrated that they cannot be trusted to use emerging technology responsibly,” said Katherine Gallagher, Senior Staff Attorney at the Center for Constitutional Rights. “All parties in this case are bound by international law, including the principles of distinction, proportionality and necessity aimed at protecting civilians and civilian infrastructure in times of armed conflict, and U.S. criminal law that punishes those who commit or further war crimes.”

“The challenge before us is not whether the law can keep pace with technology, as new technologies do not create new exceptions to the law. As AI becomes increasingly embedded in military operations, governments and companies remain bound by longstanding legal obligations to protect civilians and uphold international humanitarian law and human rights,” said Michael De Dora, U.S. Advocacy Manager at Access Now. “Accountability cannot and should not be automated away. AI may accelerate decision-making, but it cannot diminish legal responsibility.”

Anthropic originally brought forth their complaint in March against the U.S. Department of War (“DOW”) to challenge the Pentagon’s decision to designate Anthropic as a national security “supply chain risk” following the company’s refusal to remove its longstanding usage restrictions against lethal autonomous weapons. However, as Amici argued in March, they remind the Court again today: a prohibition on full autonomy does not go far enough to meet the current legal constraints it should operate under according to international and domestic law.

“The court has an important role to signal what is at stake: AI-facilitated military targeting is a threat to the rule of law and to humanity. We must stop it here and now, if we don’t want it to become the blueprint for armies and conflict globally,” said Lydia de Leeuw, Strategic Litigation Lead at SOMO.

A copy of the submitted brief can be found HERE.

#####

TECH JUSTICE LAW (“TJL”) is a pioneering strategic litigation and advocacy organization bringing justice to communities harmed by tech products. TJL co-filed the first-ever, groundbreaking lawsuits against a popular, “AI” chatbot product developed by Character AI, with support by Google, raising public awareness of chatbots’ real-world harms. TJL’s cases and advocacy have also focused government attention on harmful AI products, including unlicensed therapy chatbots. TJL brings together legal experts, policy advocates, digital rights organizations, and technologists to ensure that our legal protections are fit for the digital age.

Abolitionist Law Center (“ALC”) is a nonprofit legal organization and community organizing project fighting to dismantle state and corporate violence through legal action, research, public education, and advocacy on the local, state, national, and international levels. Since 2023, ALC has investigated and reported on the emerging role of artificial intelligence in state-sponsored violence and oppression.

Access Now is an international non-profit organization and US-registered 501(c)3 tax-exempt organization working to strengthen the digital rights of people and communities at risk. Access Now routinely engages courts in the United States and abroad, advocating for privacy, freedom of expression, and civic space in the digital age.

Center for Constitutional Rights (“CCR”) is a national not-for-profit legal, educational, and advocacy organization dedicated to advancing and protecting the rights guaranteed by the United States Constitution and international law. CCR has been responsible for some of the most significant advancements in the recognition of international law in federal courts over the last four decades, and has sought to hold state and nonstate actors liable for torture, extrajudicial killings, war crimes, and crimes against humanity. CCR has an interest in the proper interpretation of international law as it applies to the use of evolving technologies in the context of armed conflict.

Centre for Research on Multinational Corporations (“SOMO”) is a non-for-profit organisation registered in the Netherlands. SOMO investigates the impacts and enablers of unjustified corporate power, with a clear goal – a fair and sustainable world in which public interests outweigh corporate interests. Headquartered in Amsterdam, SOMO works with hundreds of organisations worldwide, acting as a knowledge, research and communications hub.