SRINAGAR: A growing confrontation between the United States Department of Defence and artificial intelligence company Anthropic has exposed an uncomfortable question facing the technology industry: can AI remain ethical when governments demand strategic advantage in an increasingly volatile geopolitical environment?
The dispute centres on the use of Anthropic’s large language model, Claude, within US military operations. What began as a disagreement over ethical safeguards has now escalated into a high-stakes confrontation between state authority and corporate policy, raising concerns that political pressure may compel AI developers to dilute or abandon ethical boundaries.
Anthropic, a San Francisco-based AI company, had incorporated strict guidelines in its “acceptable use policy” governing the deployment of its models in defence operations. These restrictions prohibited the use of its AI systems in mass surveillance of United States citizens and banned their integration into fully autonomous weapons systems capable of selecting and engaging targets without human intervention.
However, the US Department of Defence has pushed back strongly against such limits. Defence Secretary Pete Hegseth recently issued an ultimatum demanding that the company relax its ethical restrictions or risk the government invoking the Defence Production Act of 1950, which would allow federal authorities to compel access to the technology.
The government has also indicated that Anthropic could be labelled a supply chain risk, potentially jeopardising its lucrative defence contracts.
The confrontation intensified after media reports alleged that Anthropic’s technology had been used during a US military operation involving the capture of former Venezuelan president Nicolás Maduro in January 2026. Although Anthropic denied that its systems had violated internal policies during the operation, the episode triggered internal debate within the company and intensified scrutiny over how military institutions deploy AI.
At the centre of the debate is the definition of “responsible AI.” Anthropic has maintained that responsible use must include clear ethical red lines, particularly regarding surveillance and autonomous weapons.
The Pentagon, however, has argued that such limitations are impractical in an era of shifting geopolitical threats. In a January memorandum, the defence secretary stated that artificial intelligence used by the military should permit “any lawful use,” suggesting that ethical constraints embedded in AI systems could restrict operational effectiveness.
The memorandum also criticised what it described as ideological “tuning” in AI models, instructing defence agencies to ensure that future AI contracts incorporate the “any lawful use” standard.
The dispute highlights a broader transformation taking place across the global AI industry. Several major technology companies have recently softened their ethical commitments as competition for defence and government contracts intensifies.
Technology firms that once emphasised strict ethical frameworks are increasingly revising those principles. Google recently updated its AI policies, removing earlier pledges that limited the technology’s use in weapons development and surveillance. OpenAI has revised its mission language, placing less emphasis on safety commitments, while Elon Musk’s xAI has accepted the Pentagon’s “any lawful use” framework for defence applications.
The changes suggest that ethical language often collides with strategic and commercial realities.
Anthropic itself appears to be reconsidering some of its commitments. In February 2026, the company updated its “responsible scaling policy,” which governs how it releases advanced AI systems. The revised policy removed a previous promise that the company would withhold powerful AI models if it could not guarantee adequate risk mitigation measures.
Company executives have acknowledged the pressure created by intense technological competition. Anthropic’s chief science officer, Jared Kaplan, noted that rapid advancements in AI meant it was difficult for any single company to maintain unilateral commitments if competitors moved ahead with fewer restrictions.
The growing role of AI in military operations has turned the technology into a strategic asset comparable to nuclear technology during the Cold War. Governments view advanced AI systems as tools for intelligence analysis, cyber operations, battlefield decision-making and potentially autonomous weapons development.
This geopolitical context is placing technology companies under increasing pressure to prioritise national security objectives over ethical safeguards.
The Anthropic dispute has emerged at a moment when global debate over military AI is intensifying. Several countries recently gathered to discuss frameworks for “responsible AI” in warfare, while the United Nations is preparing further discussions on limiting the development of lethal autonomous weapons systems.
Yet the rapid integration of AI into defence operations suggests that ethical agreements may struggle to keep pace with technological and strategic developments.
For AI companies, the dilemma is becoming increasingly clear: maintaining strict ethical limits may protect long-term societal interests but could also mean losing contracts, market share and influence in a rapidly expanding defence technology sector.
The clash between Anthropic and the US defence establishment, therefore, represents more than a corporate dispute. It signals a broader shift in which political priorities and military competition may ultimately determine how artificial intelligence is designed, deployed and regulated.
As governments deepen their reliance on AI for security and warfare, the challenge for the technology industry will be whether ethical principles can survive in a landscape increasingly shaped by geopolitical power.
