The Era of AI as Military Advisor

On June 16, 2026, MIT Technology Review released an exclusive eBook that extensively examines the use of AI models in military decision-making, including the current state of affairs. The eBook, available exclusively to subscribers, compiles six articles published by the magazine between April 2025 and April 2026, updated to reflect the latest developments. The articles were authored by James O’Donnell, a writer for the publication.

The collection of stories highlights various aspects of generative AI’s penetration into the military domain. Specific topics include:

Generative AI is learning to spy for the US military — Explores the potential application of generative AI in intelligence activities. By analyzing vast quantities of open-source information and signals intelligence, generative AI is suggested to either replace or complement traditional human analysis.

Phase two of military AI has arrived — Covers the transition from experimental usage to operational deployment. The article describes how AI is progressing beyond tactical decision-making and data analysis to becoming embedded in decision-making processes for entire operations.

A defense official reveals how AI chatbots could be used for targeting decisions — One of the most noteworthy stories in the eBook. It discusses the examination of AI chatbot interfaces for presenting target information and assisting human decision-making. While ethical and legal issues are raised, discussions about implementation are advancing.

The Pentagon is planning for AI companies to train on classified data, defense official says — Explores the concept of using classified information to train AI models. Historically, there has been a significant gap between military intelligence and AI development, but efforts to bridge this divide are rapidly accelerating. Balancing data security with model reliability is a key focus.

How AI is turning the Iran conflict into theater — Analyzes the role of AI in information manipulation and propaganda. The article portrays how conflicts are being orchestrated and reconstructed in real time through AI technologies.

Additionally, the eBook includes a related story titled “10 Things That Matter in AI Right Now: The New War Room,” which provides a comprehensive overview of critical topics in AI and situates military AI within this broader context.

Content and Significance of the Included Articles

The primary significance of this eBook lies in its presentation of military AI as an interconnected series of phenomena rather than isolated topics. Intelligence activities, targeting decisions, training on classified data, and manipulation of conflict narratives—these aspects have traditionally been discussed separately. However, the rapid development of generative AI is integrating these elements, leading to a transformation of military doctrines.

The potential use of chatbots to assist in targeting decisions is particularly striking, as it indicates that the once science-fiction notion of “machine-assisted decisions on the use of lethal force” is becoming increasingly plausible. Although the principle of human oversight in decision-making (Human-in-the-loop) is expected to remain intact, the sheer quantity and quality of information provided by AI could significantly influence human judgment, raising concerns about the risks of bias and over-reliance.

Plans to train AI models using classified data also represent a major technological and institutional turning point. Until now, military AI has relied on public or simulated data for training. By integrating real battlefield and intelligence data into model training, the accuracy and effectiveness of AI systems could be dramatically enhanced. However, this approach introduces new challenges, such as the risk of data leaks and increased vulnerability to adversarial attacks.

Transformations in Military Decision-Making

Brought by AI

The overarching narrative from this eBook suggests that AI is evolving from a mere auxiliary tool to a substantial “advisor” in military operations. Traditionally, military AI has been applied to limited tasks such as reconnaissance through image recognition or optimizing logistics. The advent of generative AI, however, enables sophisticated support for complex mission planning and risk assessment through natural language dialogue.

From the specific stories, the following trends emerge:

1. Efforts to make AI decision-making more “explainable” are underway. By adopting chatbot interfaces, humans can better understand the rationale behind AI-recommended targets.
2. The expansion of training data into classified domains is increasing both the performance and risks associated with AI models.
3. AI-driven information warfare is becoming increasingly intertwined with physical combat situations, fundamentally altering the nature of conflicts.

Editorial Opinion

In the short term, two themes from the eBook stand out as likely to manifest as concrete projects between late 2026 and 2027: the use of AI chatbots for targeting decisions and plans to train AI on classified data. The U.S. Department of Defense is reportedly already running several pilot programs in these areas, with budget discussions expected to take place in Congress soon. These developments are likely to extend their influence to NATO countries, Japan, and other allied nations, shaping their respective military AI strategies.

In the long term, as AI solidifies its role as a military advisor, the very nature of warfare could be transformed. Decision-making speeds may surpass human cognitive limits, possibly leading to an era of “algorithmic warfare,” where strategic interactions occur directly between AI systems. However, this also raises the specter of AI-driven errors resulting in war crimes or civilian casualties. While the establishment of international regulatory frameworks is urgently needed, the conflicting interests of nations are already complicating such efforts.

As an editorial team, we pose the question: What happens when technological advancements outpace the ethical and legal frameworks meant to govern them? If AI-assisted targeting decisions become a reality, who will bear responsibility for mistakes? Should it be the AI developers, the commanding officers implementing the AI’s recommendations, or the ultimate decision-makers? Current international humanitarian law does not account for such scenarios, nor are there precedents to guide judgment. It is imperative for military AI experts and international legal scholars to engage in detailed discussions based on concrete scenarios.

References

• MIT Technology Review: Exclusive eBook: How AI is becoming the next military advisor — Published June 16, 2026