Current Affairs Topics Archive
International Relations Economics Polity & Governance Environment & Ecology Science & Technology Internal Security Geography Social Issues Art & Culture Modern History

US Treasury says stopping use of Anthropic's tech

March 03, 2026 5 min read Science & Technology International Relations GS2 GS3
On This Page ▼

What Happened

  • US President Donald Trump directed every federal agency to immediately cease use of Anthropic's AI technology, including its Claude large language model; the US Treasury Department became one of the first agencies to formally confirm compliance with the order.
  • The dispute originated from a Pentagon demand that Anthropic agree to give the US military unrestricted access to its AI models — including for autonomous weapons and domestic surveillance — which Anthropic CEO Dario Amodei refused, citing the company's core safety commitments.
  • Defense Secretary Pete Hegseth subsequently designated Anthropic a "supply chain risk," prohibiting any contractor, supplier, or partner doing business with the US military from conducting any commercial activity with Anthropic.
  • In parallel, OpenAI — Anthropic's chief commercial rival — quickly struck a deal with the Pentagon and positioned itself to absorb federal contracts vacated by the ban.
  • By late March 2026, a federal judge in San Francisco (Judge Rita F. Lin) issued a preliminary injunction temporarily blocking the Pentagon's designation of Anthropic as a supply chain risk and halting Trump's order against federal use of Anthropic's technology.

Static Topic Bridges

AI Governance and the Debate Over Safety vs. Capability in Defense Applications

AI governance refers to the policies, regulations, and institutional frameworks that guide the development, deployment, and oversight of artificial intelligence systems. A central tension in AI governance is between maximizing AI capability for competitive or security advantage versus maintaining human oversight, ethical guardrails, and safety restrictions on harmful applications.

  • Anthropic's "responsible scaling policy" commits the company to not deploying AI for autonomous lethal weapons or mass surveillance without meaningful human oversight — guardrails the Pentagon sought to remove.
  • OpenAI's amended Pentagon deal added explicit language prohibiting the AI system from being "intentionally used for domestic surveillance of U.S. persons and nationals" — a clause added after employee backlash and public criticism.
  • The EU AI Act (2024) categorizes AI systems used in critical infrastructure, law enforcement, and migration management as "high risk," requiring conformity assessments and human oversight.
  • OECD AI Principles (2019) and the Bletchley Declaration (2023) emphasize human-centric, safe, transparent AI — frameworks the US government's military AI acquisition posture appears to be moving away from.

Connection to this news: The Anthropic ban illustrates the global tension between state actors seeking unconstrained AI military tools and technology companies asserting safety limits — a debate that India's own AI governance framework must navigate as it develops IndiaAI Mission applications.

Dual-Use Technology — AI at the Intersection of Commerce and Defense

Dual-use technology refers to technologies, software, or knowledge that have both civilian and military applications. Large language models (LLMs) are quintessentially dual-use: the same model that helps write code or analyze documents can be used for surveillance, autonomous targeting, propaganda generation, or cyber operations.

  • The US Export Administration Regulations (EAR) and International Traffic in Arms Regulations (ITAR) are the primary US laws governing the export of dual-use and defense technologies; AI software is increasingly falling under EAR scrutiny.
  • WASSENAAR Arrangement (1996) is a multilateral export control regime covering conventional arms and dual-use goods; the addition of AI and cybersecurity tools to WASSENAAR control lists has been debated since 2021.
  • India's IndiaAI Mission (launched March 2024, outlay ₹10,372 crore) specifically includes defense AI applications, with DRDO's Evaluating Trustworthy Artificial Intelligence (ETAI) Framework (October 2024) establishing risk-based criteria for AI in the Indian armed forces.
  • China's "military-civil fusion" strategy explicitly directs commercial AI companies to contribute to military AI development — a model the US and India are watching closely.

Connection to this news: The Trump administration's pressure on Anthropic reflects a broader US government push to mobilize commercial AI for military ends; India faces analogous questions about whether and how to direct its domestic AI industry toward defense applications under the IndiaAI Mission.

India's AI Policy Landscape — IndiaAI Mission and Governance

India launched the IndiaAI Mission in March 2024 with a cabinet outlay of ₹10,372 crore (approximately $1.25 billion) over five years, aiming to build India's domestic AI ecosystem, computing infrastructure, and talent base.

  • The mission's seven governance principles include Trust, People First, Innovation over Restraint, Fairness & Equity, Accountability, Understandable by Design, and Safety, Resilience & Sustainability.
  • The government has committed to onboarding more than 38,000 GPUs to a shared compute facility accessible to Indian startups and academic institutions at subsidized rates.
  • India does not yet have a comprehensive AI regulation law (unlike the EU AI Act), operating instead through sector-specific guidelines and the DPDP Act 2023 for data aspects.
  • The IndiaAI Mission shortlisted 12 teams for development of indigenous AI models (large language models) in 2025, aiming to reduce dependence on US and Chinese AI platforms.
  • India's AI Safety Institute is being established under the mission as the national institutional mechanism for AI risk assessment.

Connection to this news: The US federal AI market dynamics — with Anthropic losing government contracts and OpenAI gaining them — directly shapes the competitive landscape for global AI firms operating in India; it also provides India a case study in the risks of over-dependence on foreign AI platforms for sensitive government functions.

Key Facts & Data

  • Trump directed all US federal agencies to stop using Anthropic technology; US Treasury Secretary Scott Bessent confirmed compliance.
  • Pentagon designated Anthropic a "supply chain risk" under Defense Secretary Pete Hegseth.
  • Anthropic CEO Dario Amodei refused Pentagon demands for unrestricted access, citing refusal to enable autonomous weapons and domestic surveillance.
  • A federal judge (Judge Rita F. Lin, San Francisco) issued a preliminary injunction in late March 2026, temporarily halting both the supply chain risk designation and Trump's ban.
  • OpenAI amended its Pentagon deal to explicitly bar "domestic surveillance of U.S. persons and nationals."
  • IndiaAI Mission: launched March 2024, outlay ₹10,372 crore, 38,000+ GPUs onboarded.
  • India's ETAI Framework for defense AI: launched October 17, 2024.