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

OpenAI strikes Pentagon deal with 'safeguards' as Trump dumps Anthropic

February 28, 2026 5 min read Science & Technology International Relations GS3
On This Page ▼

What Happened

  • The Trump administration ordered all US federal agencies to cease using services of AI company Anthropic, with Defense Secretary Pete Hegseth designating Anthropic a "Supply-Chain Risk to National Security" — a designation normally reserved for foreign adversaries such as Chinese telecom firms.
  • The ban followed a dispute in which Anthropic sought contractual guarantees from the Pentagon that its AI systems (Claude) would not be used for domestic mass surveillance of Americans or for fully autonomous weapons systems.
  • Anthropic stated it objects to autonomous weapons use because "today's frontier AI models are not reliable enough for fully autonomous weapons" and that mass domestic surveillance "constitutes a violation of fundamental rights."
  • Within hours of the Anthropic ban, OpenAI announced a deal with the Pentagon to deploy its models on classified military networks, with CEO Sam Altman claiming OpenAI's agreement includes shared prohibitions on domestic mass surveillance and human-supervised use of force.
  • The episode marks a significant flashpoint in the governance of AI in military applications and the political dynamics surrounding AI companies' ethical constraints.

Static Topic Bridges

Lethal Autonomous Weapons Systems (LAWS): International Governance Debate

Lethal Autonomous Weapons Systems (LAWS) — commonly called "killer robots" — are weapons that can identify and engage targets without direct human control. The international community has debated regulation of LAWS through the Convention on Certain Conventional Weapons (CCW) at the UN since 2014. In December 2024, the UN General Assembly adopted a resolution on LAWS with 166 votes in favour (3 against, 15 abstentions), calling for a legally binding instrument by 2026. The central ethical concern is the "accountability vacuum" — if an autonomous weapon causes unlawful harm, it is unclear whether the programmer, manufacturer, commander, or state bears legal responsibility under international humanitarian law (IHL).

  • LAWS: Weapons that select and engage targets autonomously, without human control
  • UN CCW: Convention on Certain Conventional Weapons; LAWS discussed in Group of Governmental Experts (GGE) since 2014
  • UNGA Resolution on LAWS: December 2, 2024; 166 in favour, 3 against, 15 abstentions
  • UN Secretary-General recommendation: Legally binding instrument on LAWS by 2026
  • Countries calling for ban: ~30 states + 165 NGOs (include Austria, Chile, New Zealand)
  • US position: Does not support a categorical ban; argues LAWS can improve targeting accuracy
  • IHL principles at stake: Distinction (combatants vs. civilians), Proportionality, Precaution
  • "Accountability vacuum": Legal gap when AI system — not a human — makes targeting decision

Connection to this news: Anthropic's specific objection to its AI being used in "fully autonomous weapons" directly invokes the LAWS governance debate — a GS3 science and security question that is increasingly Mains-relevant.

AI Ethics, Responsible AI, and Corporate Governance Frameworks

The Anthropic-Pentagon dispute illustrates the concept of Responsible AI (RAI) — a framework that governs the development and deployment of AI systems with reference to principles of safety, transparency, accountability, fairness, and human oversight. Major AI governance frameworks include the OECD Principles on AI (2019), the EU Artificial Intelligence Act (2024 — world's first comprehensive AI regulation), and India's National Strategy for Artificial Intelligence (NITI Aayog, 2018). The US Executive Order on AI (October 2023, Biden administration) required federal agencies to conduct risk assessments before deploying AI in sensitive applications — an order the Trump administration subsequently revised.

  • OECD Principles on AI (2019): Foundational international soft-law AI governance framework; India is an adherent
  • EU AI Act (2024): Risk-based regulation; bans certain uses (social scoring, real-time biometric surveillance); high-risk AI requires conformity assessments
  • India's AI strategy: NITI Aayog — "Responsible AI for All" (2021 Principles document); MeitY leading national AI Mission (IndiaAI)
  • IndiaAI Mission: Approved February 2024; ₹10,371 crore; focus on compute, datasets, AI startups
  • Anthropic's usage policy: Prohibits use of Claude for autonomous weapons and mass surveillance
  • "Supply-Chain Risk" designation: US legal mechanism (typically used against Huawei, ZTE) — unprecedented application to a domestic AI firm

Connection to this news: The Trump administration's use of a national security supply-chain tool against a domestic AI company for maintaining ethical guardrails reveals the tension between commercial AI governance frameworks and state security imperatives — directly testable in GS3 S&T policy.

US-China AI Competition and National Security Framing of AI

The framing of AI as a national security asset — and the "supply-chain risk" designation — reflects the broader US-China strategic competition over AI dominance. The US has imposed semiconductor export controls (October 2022, expanded 2023 and 2024) on China to restrict its access to advanced chips needed for AI training. China's AI firms (Baidu, Huawei, DeepSeek) have been designated security risks by various US agencies. The OpenAI-Pentagon deal — deploying AI on classified military networks — represents a deepening of the US military's AI integration strategy. India is navigating this competition cautiously, seeking technology partnerships with the US (iCET: Initiative on Critical and Emerging Technologies, 2023) while maintaining strategic autonomy.

  • US semiconductor export controls on China: October 2022 (BIS rule); expanded 2023 and 2024
  • iCET: Initiative on Critical and Emerging Technologies (India-US); launched January 2023
  • China's AI firms designated security risks by US: Huawei (banned from US government networks, 2019)
  • OpenAI Pentagon deal: Deploy models on classified DoD networks (February 27, 2026)
  • Anthropic designation: "Supply-Chain Risk to National Security" (Defense Secretary Hegseth, February 2026)
  • Pentagon AI budget: DoD requested $1.8 billion for AI (FY2025)
  • DeepSeek R1 release (January 2025): Open-source Chinese AI model matched US capabilities — demonstrated competitive pressure

Connection to this news: The speed of the OpenAI-Pentagon deal following the Anthropic ban illustrates how AI governance is being subordinated to geopolitical competition — a GS2-GS3 intersection on technology diplomacy.

Key Facts & Data

  • Anthropic banned by Trump: February 27, 2026 — "Supply-Chain Risk to National Security" designation
  • OpenAI Pentagon deal: Announced February 27, 2026 (within hours of Anthropic ban)
  • Pentagon contract contested: Worth up to $200 million (Anthropic)
  • Anthropic's objections: No autonomous weapons; no domestic mass surveillance
  • UN LAWS resolution: December 2, 2024; 166 in favour
  • UNGA target for binding LAWS instrument: By 2026
  • EU AI Act: Adopted 2024; world's first comprehensive AI regulation
  • OECD AI Principles: 2019 — India is signatory
  • IndiaAI Mission: ₹10,371 crore approved February 2024
  • iCET (India-US): Launched January 2023
  • US semiconductor export controls on China: October 2022 (BIS Entity List expansion)