Can AI systems make moral decisions in war?
Military forces worldwide are increasingly integrating artificial intelligence into combat operations, raising questions about whether AI systems possess the moral capacity to make decisions in warfare. Global military spending on AI applications is projected to grow significantly, with systems like Palantir's Maven Smart System already deployed in active conflicts.
Artificial intelligence is fundamentally transforming military operations across multiple conflict zones. Autonomous systems now operate in Ukraine, Gaza, and Iran, replacing traditional decision-making structures with machine-driven processes. The integration of AI into what military strategists call the kill chain—the sequence from target identification to strike execution—has compressed decision timelines and reduced human oversight.
The scale of this technological shift is reflected in defence budgets. Global military expenditure reached $2.87 trillion last year, with AI-specific military spending accelerating rapidly. Industry projections estimate growth from $11.7 billion currently to nearly $19.3 billion by 2030. The Trump administration has proposed $1.5 trillion in defence spending for the upcoming fiscal year, explicitly aimed at transitioning from traditional military-industrial structures toward military-technology integration.
A central focus of this transformation is the OODA loop—a decision-making framework developed by US Air Force Colonel John Boyd that prioritizes speed over certainty. By compressing observation, orientation, decision, and action into rapid cycles, military commanders can respond to battlefield changes faster than adversaries. Project Maven, Palantir's AI intelligence platform paired with Anthropic's Claude, exemplifies this approach by providing real-time battlefield analysis.
However, recent incidents highlight the risks. During the first day of US-Israel operations against Iran in February, Maven-assisted strikes targeted over 1,000 locations, including a primary school in Minab where more than 150 people died, predominantly children. Pentagon investigations into the incident remain incomplete, with reports indicating the error stemmed from outdated satellite intelligence. This case underscores a critical gap: while AI systems excel at pattern recognition and speed, they lack mechanisms for moral evaluation of their targeting decisions.
The underlying logic of AI-driven warfare is not new. During drone operations in the 2000s and 2010s, algorithmic systems flagged individuals matching suspicious behavioral profiles, generating computational kill lists. What has changed is scale, speed, and integration. Modern systems process vastly more data and execute decisions with minimal human intervention, raising fundamental questions about accountability and the capacity of machines to navigate moral complexity in warfare.
Bell tracks these organizations in depth — profiles, people, signals, and history. See them inside Bell →
Enter the market with the full picture.