West Pointers Can Be Trained to Better Evaluate, Appreciate AI, Study Finds

Why It Matters

Reducing automation bias among future officers can improve decision‑making safety and operational effectiveness, while the training model offers a roadmap for broader AI literacy initiatives.

Key Takeaways

• West Point cadets scored twice the AI knowledge of public
• Cadets half as likely to trust erroneous chatbot answers
• Training emphasized checking AI confidence indicators, unlike typical users
• 88% of cadets view AI as beneficial versus 73% of public

Pulse Analysis

Trust in artificial intelligence remains uneven, with Americans often more skeptical than peers in China, yet many still accept flawed chatbot responses. This gap poses a risk for sectors where rapid, accurate decisions are critical, especially the military, which is integrating AI‑driven decision‑support systems to shrink sensor‑to‑shooter cycles. Understanding concepts like automation bias—the tendency to over‑rely on automated outputs—and algorithm aversion is essential for leaders who must balance speed with judgment.

The recent paper, authored by scholars from Georgetown, the University of Pennsylvania and the U.S. Military Academy, surveyed 236 West Point cadets against a demographically matched sample of 702 civilians. Cadets demonstrated AI knowledge scores nearly double those of the public and were less than 50% as likely to accept a chatbot’s wrong answer. Crucially, they habitually examined the AI tool’s confidence indicators, a practice seldom seen among ordinary users. This disciplined approach aligns with West Point’s curriculum theme “The Human and the Machine,” which emphasizes calibrated confidence—matching expectations of AI accuracy to reality.

Beyond the academy, the study offers a template for scaling AI literacy across the armed forces and even the civilian workforce. If similar training can be embedded in broader military education, it could mitigate the national‑security concerns tied to declining public trust in AI. Moreover, the data suggest that fostering optimism—87.7% of cadets see strong beneficial applications—while teaching critical evaluation may bridge the trust gap, preparing both soldiers and citizens for an increasingly AI‑augmented world.