Sources Allege Military Used Anthropic AI in Iran Operations Amid Federal Ban and Industry Backlash

Illustration for: Sources Allege Military Used Anthropic AI in Iran Operations Amid Federal Ban and Industry Backlash
AI-generated illustration. Visual interpretation does not represent real individuals or scenes.

THE BARE STORY

President Trump has ordered federal agencies to cease using technology from AI company Anthropic, granting a six-month phase-out period, while Defense Secretary Pete Hegseth has designated the firm a "supply chain risk." Despite this directive, sources familiar with the matter stated that the U.S. military utilized Anthropic’s Claude AI model during attacks on Iran over the weekend and continues to use the tool.

The government prohibition follows a breakdown in negotiations regarding the terms of military application. Anthropic CEO Dario Amodei reportedly insisted on "guardrails" explicitly preventing the military from using the AI for mass surveillance on Americans or to power fully autonomous weapons. In response, the Pentagon demanded the ability to use the technology for "all lawful purposes." Pentagon officials, including Chief Technology Officer Emil Michael, argued that mass surveillance is already illegal and that internal policies regulate autonomous weapons.

The dispute has triggered a mobilization of workers across the technology sector. Employees at companies including Alphabet, OpenAI, and Salesforce have circulated letters and petitions supporting Anthropic’s position or demanding stricter limits on military AI usage. One open letter, signed by nearly 900 tech workers, accused the government of using fear to divide the industry, while another petition urged Congress to review the "supply chain risk" designation.

Internal dissent has also been reported at Google, where over 100 employees signed a memo asking management to establish ethical boundaries similar to those proposed by Anthropic. Jeff Dean, Google’s chief scientist, publicly acknowledged these concerns, noting that mass surveillance raises significant constitutional and political issues.

Same Facts. Different Perspectives.

Two AI models. Two viewpoints. One factual foundation.

• Weaponizing Administrative Retribution Prioritizing the protection of dissent, this lens views the "supply chain risk" designation by Secretary Hegseth not as a technical assessment, but as a punitive political tactic. By phasing out Anthropic following failed negotiations, the administration appears to be weaponizing procurement power to coerce private entities into abandoning ethical standards. The logic holds that labeling a vendor a risk simply for refusing to enable mass surveillance sets a dangerous precedent for government accountability.

• Codifying Civil Liberties Valuing proactive constraints on state power, this perspective argues that the Pentagon's assurance that "mass surveillance is already illegal" is insufficient without binding technological hard-stops. Anthropic’s demand for explicit "guardrails" against autonomous weapons and surveillance is seen as a necessary firewall against potential abuse. Relying solely on "internal policies" or current interpretations of "lawful purposes" is viewed as naive given the historical tendency of security agencies to expand their reach during conflicts.

• Mobilizing The Silicon Conscience Viewing corporate power as a potential check on government overreach, this camp champions the mobilization of workers at Alphabet, OpenAI, and Salesforce as a vital democratic mechanism. The petitions and open letters are interpreted not as insubordination, but as a "Collective Conscience" ensuring that technology serves humanity rather than just the military-industrial complex. The support from figures like Jeff Dean highlights that constitutional concerns regarding AI are valid ethical boundaries, not impediments to national security.

How it may affect me

As a U.S. reader:

• You may face uncertainty regarding the technical limits of government monitoring, as the Pentagon's refusal to accept Anthropic's specific guardrails against mass surveillance leaves civil liberty protections reliant on the interpretation of lawful purposes rather than binding software constraints.

• National defense operations could experience supply chain instability or strategic friction, as the military is currently using Anthropic's AI for active strikes on Iran despite a presidential directive to phase out the vendor within six months due to risk designations.

• The availability of advanced technology for national security may shift based on private sector labor movements, as widespread petitions by workers at major firms like Alphabet and Salesforce seek to pressure other companies into adopting similar ethical restrictions on government contracts.

Read the story at

Note: All TheBareNews content is AI-generated. For additional context, reporting, and updates, you are invited to explore the news outlets linked above.