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

Anthropic’s resistance to the U.S. Department of Defence, OpenAI’s entry: Explained

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

What Happened

  • The US Department of Defense (DoD/Pentagon) formally designated Anthropic — maker of the Claude AI — a "supply chain risk" effective immediately on March 5–6, 2026.
  • The designation followed Anthropic CEO Dario Amodei's refusal to allow Claude AI systems to be used for: (a) fully autonomous weapons systems with no human involvement in targeting/firing decisions, and (b) mass domestic surveillance of Americans.
  • The "supply chain risk" label is typically reserved for foreign adversaries and effectively bars the US government from using Anthropic's products.
  • Within hours, OpenAI signed a contract with the DoD to provide AI services in Anthropic's place — though OpenAI later had to revise its contract after backlash over autonomous weapons provisions.
  • In an ironic twist, Claude AI models were simultaneously reported to be in use by the US military for planning strikes against Iran, even as the DoD designated Anthropic a risk.

Static Topic Bridges

Lethal Autonomous Weapons Systems (LAWS) and International Governance

Lethal Autonomous Weapons Systems (LAWS) are weapons that can independently identify, select, and engage targets without direct human involvement in the killing decision. They represent one of the most contested ethical and legal frontiers in AI governance.

  • The UN Convention on Certain Conventional Weapons (CCW) Group of Governmental Experts (GGE) has been debating LAWS since 2016 without reaching a binding agreement.
  • UN General Assembly (December 2, 2024): Passed a resolution on LAWS with 166 votes in favour, 3 against (Belarus, DPRK, Russia), 15 abstentions — signalling strong global consensus for regulation.
  • "Meaningful human control": the most contested concept in LAWS debates; most states agree humans must remain in the decision loop, but no agreed definition of how much autonomy is permissible exists.
  • UN Secretary-General called for a legally binding treaty banning LAWS that operate without human control by 2026.
  • Key legal concern: accountability gap — if an autonomous weapon commits a war crime, who is responsible? The programmer? The commander? The state?

Connection to this news: Anthropic's red line — no fully autonomous lethal systems — aligns with the broader international consensus around "meaningful human control." The DoD's pressure on Anthropic represents the US government pushing against the international norm the US itself nominally supports in UN forums.

AI Ethics and Corporate Governance in AI Development

Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, and others who left OpenAI over safety concerns. The company is structured as a Public Benefit Corporation and has adopted a distinctive "responsible scaling policy" that constrains what its AI can be used for.

  • Anthropic's founding principle: build AI that is safe, beneficial, and understandable — known as "Constitutional AI" (training AI to follow a set of values).
  • "Red lines" Amodei refused to cross: mass surveillance of Americans; fully autonomous lethal targeting with no human oversight.
  • Supply chain risk designation: normally reserved for Chinese companies (Huawei, ZTE) or state actors; applying it to a US AI company is unprecedented.
  • Amodei's response: Anthropic will challenge the designation in court, calling it "legally unsound."
  • OpenAI's position: initially stated it shared Anthropic's two red-line concerns but signed the DoD contract anyway, then faced backlash and revised contract language.
  • The incident illustrates the tension between AI companies' stated ethics commitments and commercial/governmental pressure.

Connection to this news: The Anthropic-DoD conflict is the first major case of a leading AI company publicly refusing a government's demand on safety-ethical grounds and facing regulatory retaliation. For UPSC, this connects to debates about AI governance, corporate accountability, and the dual-use nature of advanced technology.

India's AI Policy and Autonomous Systems Context

India is developing its own AI governance framework and has expressed views on LAWS in international forums. The intersection of AI, defence, and ethics is relevant to India's security and technology policy.

  • NITI Aayog's National AI Strategy (2018) and the India AI Mission (2024): emphasise responsible AI development with a focus on economic applications.
  • India's defence AI: DRDO is developing AI-powered drone systems; India's position on LAWS in the CCW GGE has been cautious — not calling for a ban but supporting the human control principle.
  • Dual-use technology: AI developed for commercial purposes (image recognition, language models) can be re-purposed for military applications — a regulatory challenge India faces too.
  • India's AI regulatory approach: The Digital Personal Data Protection Act (2023) addresses data governance; a dedicated AI regulatory framework is under development.
  • India at the UN: supported the 2024 UNGA resolution on LAWS, reflecting alignment with the "meaningful human control" camp.

Connection to this news: As India develops its own defence AI capabilities and debates AI regulation, the Anthropic case provides a real-world precedent for what happens when commercial AI ethics policies clash with government security demands — a dilemma India will face as its AI sector matures.

Key Facts & Data

  • Anthropic founded: 2021 by Dario and Daniela Amodei (ex-OpenAI)
  • DoD supply chain risk designation: effective March 5–6, 2026
  • Reason: Anthropic refused AI use for fully autonomous lethal weapons and mass domestic surveillance
  • OpenAI DoD contract: signed hours after Anthropic's designation; later revised after backlash
  • Claude AI: reportedly used for US military strike planning against Iran (simultaneous with designation)
  • UN CCW GGE on LAWS: deliberating since 2016; no binding treaty yet
  • UNGA resolution on LAWS (Dec 2, 2024): 166 for, 3 against, 15 abstentions
  • "Meaningful human control": central contested concept in LAWS governance
  • Anthropic's legal status: Public Benefit Corporation (US)
  • India AI Mission (2024): $1.2 billion programme for AI infrastructure and compute capacity
  • India's position on LAWS: supports human control principle; not seeking outright ban