Maj. Gen. William “Hank” Taylor, the acting Eighth Army commander in South Korea, said he is experimenting with generative AI chatbots, telling DefenseScoop, “Chat and I are really close lately.” He added that he is “trying to build models to help all of us,” including those that support individual decision-making processes that affect readiness.
AI use in the military
Taylor told DefenseScoop that the Eighth Army is “regularly using” AI for predictive analysis in sustainment and to draft weekly intelligence reports, and he wants to expand intelligence use cases.
He framed the push within broader modernization efforts in Korea across counter-UAS, medical and aviation initiatives, while emphasizing that he has the “right equipment,” training and threat awareness to remain ready.
Download the SAN app today to stay up-to-date with Unbiased. Straight Facts™.
Point phone camera here
U.S. Special Operations Forces, for example, aim to utilize AI for tasks such as paperwork, logistics and analyzing doctrine, seeking to reduce personnel’s “cognitive burden.” According to Army Col. Rhea Pritchett, the program executive officer of SOF Digital Applications, the goal is to free up operators’ “precious time to critically think” about mission actions.
Why commanders see value in faster decisions
Business Insider reported that leaders like Taylor focus on speed — often framed as the OODA loop (observe, orient, decide, act) — to make faster, better decisions than their adversaries.
Bianca Herlory, the Joint Staff AI lead, previously said that AI can help integrate and analyze global operations for “better, faster decisions.” Former Air Force Secretary Frank Kendall argued that decisions on future battlefields may be made at “machine speed,” Business Insider reported in January.
What caution and guardrails look like
Pentagon guidance advises caution with public models, as they can expose sensitive data and yield flawed results. Hutchins Data Strategy CEO Chris Hutchins told NewsNation that AI should “empower” but not replace human judgment. He warned that off-the-shelf systems may also be available to adversaries.
Unbiased. Straight Facts.TM
In May, the Army launched a secure Enterprise LLM Workspace AI, powered by Ask Sage IL5, to handle controlled unclassified information.
“Trust and culture,” Hutchins said, “those things are always going to be a factor, particularly when you’re talking about chain-of-command.”
In May, the Army launched its Enterprise Large Language Model Workspace for simple text tasks to “streamline communication, enhance operational efficiency and drive innovation.” Ask Sage IL5 powers the model, which holds accreditation to handle sensitive but unclassified government information.
In August, Army Chief Information Officer Leonel Garciga asked whether AI is “worth the squeeze” compared with less “cool” tools, according to DefenseScoop.
“Our big lesson learned is,” Garciga said, “this is expensive stuff to do. I think that’s really driving our peeling back and tightening the guardrails on use cases.”
Ars Technica pointed to the 2023 State Department “best practices” declaration, which stressed human control and the ability to disengage systems that behave unexpectedly.
Where this goes next
Taylor said generative AI is “very, very interesting” to him as a commander. He is keen to apply it to leadership and decision-making models, while the Army continues testing AI for logistics, reporting and other back-office tasks.
He indicated interest in expanding intelligence applications as services refine policies that stress security and output verification.
