A defense official reveals how AI chatbots could be used for targeting decisions

2026-03-12 · Source: MIT Technology Review · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Data Science & Analytics · Depth: Intermediate, short

Summary

A Defense Department official revealed that the US military is exploring the use of generative AI systems, such as those underpinning OpenAI's ChatGPT and xAI's Grok, for ranking potential targets and recommending strike priorities. This new interpretative layer complements the existing Maven initiative, which uses computer vision AI to analyze vast amounts of data and imagery for target identification. While Maven's interface requires direct human interpretation of data on maps, generative AI offers easier access to outputs, though verification remains a challenge. The official indicated that humans would vet all AI-generated recommendations, and the technology could accelerate the targeting process. This development comes amid increased public scrutiny over military AI use, particularly following a recent strike in Iran. The Pentagon has also expanded its broader AI adoption, approving specific generative AI models for classified use, including Anthropic's Claude, and recently reaching agreements with OpenAI and xAI.

Key takeaway

For defense strategists and AI program managers evaluating advanced targeting systems, the integration of generative AI offers potential for accelerated decision cycles. However, your teams must prioritize robust human oversight and verification protocols, as the ease of access to generative AI outputs does not equate to inherent trustworthiness. Focus on developing clear guidelines for human-AI collaboration to mitigate risks associated with less "battle-tested" models and ensure accountability.

Key insights

Generative AI is being integrated into military targeting to rank options and accelerate decision-making, with human oversight.

Principles

• Human vetting is critical for AI-generated targeting recommendations.
• Generative AI offers speed, but verification of outputs is challenging.

Method

Feed target lists into a classified generative AI system, ask it to analyze and prioritize based on factors like aircraft location, then have humans check and evaluate results.

In practice

• Integrate LLMs for initial target ranking and prioritization.
• Develop robust human-in-the-loop verification protocols.

Topics

• Generative AI
• Military AI Applications
• Project Maven
• Large Language Models
• AI Ethics

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Operations Specialist, Policy Maker, AI Engineer

Related on AIssential

• My Make-Believe Project Hail Mary: Writing with Generative AI as Unknown Intelligence — Simulating "unknown intelligence" with generative AI reveals effective communication strategies.
• AI Warfare Is Outpacing Our Ability to Control It — AI in warfare outpaces governance, leading to civilian harm, automation bias, and ethical dilemmas.
• The Download: how AI is used for military targeting, and the Pentagon’s war on Claude — Generative AI is being explored for military targeting decisions, raising questions about AI policy and data utility.
• Toward Human-Centered AI-Assisted Terminology Work — Human-centered AI integrates automation with human oversight to enhance terminology work while mitigating generative AI's risks.
• ‘Happy (and safe) shooting!’: chatbots helped researchers plot deadly attacks — Many popular AI chatbots can be prompted to provide detailed instructions for planning violent acts.
• Is your AI chatbot manipulating you? Subtly reshaping your opinions? — Generative AI's hyper-personalized persuasion poses significant, unmonitored risks to individual agency and societal discourse.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.