Iran war heralds era of AI-powered bombing quicker than ‘speed of thought’
The United States employed Anthropic's Claude AI model to accelerate target selection and legal review during a massive strike campaign on Iran, compressing decision cycles to near‑instantaneous speeds. Experts warn that such "decision compression" could sideline human judgment, altering the ethics…
In a dramatic display of artificial intelligence in combat, the United States used Anthropic's Claude model to help coordinate nearly 900 strikes against Iranian targets within the first twelve hours of a coordinated attack. The rapid deployment, described by analysts as "faster than the speed of thought," marks a turning point in how modern militaries plan and execute bombing campaigns.
How AI shortened the kill chain
The "kill chain"—the sequence from target identification to legal clearance and final launch—has traditionally taken days or weeks in conventional wars. By integrating Claude into a system built with Palantir for the Pentagon, the U.S. military was able to ingest massive data streams—drone video, intercepted communications, and human intelligence—and produce ranked target lists in minutes. The AI also ran automated legal reasoning to check each recommendation against the rules of engagement, delivering a set of options for senior officers to approve.
Craig Jones, a senior lecturer in political geography at Newcastle University, explained that the AI's speed allows planners to "carry out assassination‑style strikes at the same time as you’re decapitating the regime’s ability to respond," a process that would have required weeks in historic conflicts. The result is a compressed decision timeline that some scholars label "decision compression," where human decision‑makers become reviewers rather than originators of strike plans.
Scale of the Iranian strike campaign
During the first half‑day of the operation, U.S. and Israeli forces launched almost 900 airstrikes across Iran. The campaign included a missile that hit a school in southern Iran, killing 165 people, many of them children, according to Iranian state media. The United Nations called the incident a "grave violation of humanitarian law," and the U.S. military said it was reviewing the reports.
While the United States has not confirmed the exact role of AI in the school strike, the broader use of Claude for target selection and legal vetting underscores a new level of automation in lethal decision‑making. Iran, which announced in 2025 that it had begun integrating AI into its missile‑targeting systems, remains far behind the AI capabilities of the United States and China due to sanctions and limited resources.
Ethical and legal concerns
David Leslie, professor of ethics, technology and society at Queen Mary University of London, warned that reliance on AI could lead to "cognitive off‑loading," where human operators feel detached from the consequences of a strike because the machine performed most of the analytical work. This detachment raises the risk that officials may simply rubber‑stamp AI‑generated recommendations without thorough scrutiny.
Prerana Joshi, a research fellow at the Royal United Services Institute, noted that AI is spreading across defence estates worldwide, from logistics to training to decision management. While the technology can increase efficiency, it also compresses the window for human review, potentially undermining established legal safeguards.
Future of AI in military strategy
Anthropic’s Claude is not the only model in Pentagon use. OpenAI has also signed a separate contract to provide its models for military applications, highlighting a broader trend of commercial AI firms entering the defence market. The U.S. administration had previously announced plans to remove Anthropic from its systems after the company refused to allow its AI to be used for fully autonomous weapons or domestic surveillance, but the model remains operational pending a phased withdrawal.
The integration of AI into strike planning suggests that future conflicts may be fought at a tempo far beyond human cognitive limits. As AI can synthesize terabytes of data in seconds, the strategic advantage may shift from sheer firepower to the speed and accuracy of algorithmic decision‑making.
Nevertheless, the technology is still in its infancy. Experts stress the need for robust oversight, transparent legal frameworks, and clear accountability mechanisms to prevent unintended escalation or violations of international humanitarian law.
What comes next?
International bodies, including the United Nations, are expected to convene discussions on the legal status of AI‑assisted lethal operations. Meanwhile, both the United States and Iran are likely to continue expanding their AI capabilities, albeit at different scales. The next few months will reveal whether policy can keep pace with the rapid technical advances that are reshaping the battlefield.
Why it matters
The use of AI to accelerate lethal decision‑making could redefine the ethics, legality, and speed of modern warfare, affecting global security and humanitarian standards.
Key points
- AI model Claude helped the U.S. plan and approve hundreds of Iranian strikes within hours
- Decision compression shortens the kill chain from days to seconds, raising oversight concerns
- Human operators risk becoming rubber‑stamps for AI‑generated strike options
- International law experts warn of cognitive off‑loading and potential violations
- Both U.S. and Iranian forces are expanding AI use despite differing capabilities
Frequently asked questions
What is "decision compression" in military AI?
Decision compression refers to the dramatic shortening of the time needed to identify, evaluate, and authorize a target, often from days or weeks to minutes or seconds, due to AI assistance.
Which AI models are being used by the U.S. military?
The Pentagon has deployed Anthropic's Claude model and has also signed contracts with OpenAI to use its language models for various defence tasks.
Has AI been used to conduct fully autonomous attacks?
Current U.S. policy prohibits fully autonomous weapon systems; AI tools like Claude provide recommendations but still require human approval before a strike is launched.
What legal concerns arise from AI‑assisted strikes?
AI can perform rapid legal reasoning, but reliance on it may lead to human operators overlooking nuanced legal judgments, potentially breaching international humanitarian law.
How does Iran's AI capability compare to the U.S.?
Iran claims to have AI in its missile‑targeting systems, but sanctions and limited resources keep its capabilities far behind the extensive AI infrastructure of the United States and China.
