There’s a new job on the battleground for artificial intelligence. Maj. Gen. William Taylor, commander of the United States 8th Army in South Korea, announced that he’s operating side-by-side with generative AI chatbots in hopes of boosting the ways that his headquarters makes everything from combat choices to everyday operational planning. At the Association of the U.S. Army’s annual conference in Washington, Taylor was candid regarding his partnership with AI. “Chat and I have actually become very good friends over the last little while,” he was overheard saying at a roundtable with reporters.
The general’s frank disclosure gives a sense of how Army leaders are increasingly seeking help from artificial intelligence tools to stay ahead in current warfare.
It’s more than tinkering with this and that. He’s busy attempting to create AI models capable of learning and refining decision-making processes within his command. “I’m requesting to build, attempting to build models to assist us all,” said General Taylor, who explained that such technology assists him in considering both military and personal decisions that impact thousands of troops he leads.
It’s a humble but forceful goal: make better choices, sooner. “As a commander, I’d like to make better decisions. I’d like to make sure I make decisions at the right time to give me an advantage,” Taylor said.
This focus on swiftness and timing is a central military doctrine with its own name, “OODA loop”, observe, orient, decide, and act. Leaders who transition through these phases more quickly than adversaries typically gain the upper hand over them.
Balancing AI’s Benefits and Risks in U.S. Special Operations
Taylor is not unique in welcoming AI support. Special Operations Forces of the United States also brought artificial intelligence as part of their operations, with the particular intention of “decreasing the cognitive load” on already overburdened personnel who carry heavy responsibilities.
The tools perform administration jobs, create reports, support operations planning, manage logistics, among other day-to-day operations that would otherwise occupy precious time and cognitive power.
Simple logic: if the routine can be managed by AI, then human operators can concentrate on what is best for them, strategic thinking, leadership, complex problem-solving that is dependent on human experience and judgment.
In spite of potential uses, Pentagon officials cautioned commanders experimenting with these new technologies. They are mainly concerned that generative AI can unwittingly “leak” classified information.
If not protected properly, confidential or classified data might be revealed through AI networks that aren’t developed with security levels used within the Pentagon’s office buildings.
Utilizing AI as a Force Multiplier, Not a Replacement, in Defense
Even more disconcerting is that AI is prone to highly incorrect output when it is not well-trained or monitored. Authorities point out that not having humans oversee critical wartime decisions through AI can have severe repercussions.
A U.S. Air Force test is a prime example. Attack plans were made up by AI algorithms, some 400 times quicker than human planners, a remarkable achievement that might transform warfare. But not all such plans proved implementable.
Maj. Gen. Robert Claude explained that mistakes weren’t always egregious. The AI didn’t suggest crazy things like deploying tanks on airborne operations. The mistakes were small and possibly hazardous, such as ordering the wrong kind of sensors for certain types of weather. The sort of mistakes that might well go unnoticed by someone who is not scrutinizing the AI’s advice carefully.
The U.S military’s growing relationship with artificial intelligence represents both tremendous opportunity and genuine risk. Commanders like Taylor are pioneering ways to leverage AI’s speed and analytical capabilities while navigating the technology’s limitations and potential pitfalls.
The secret would appear to lie in finding that right balance, utilizing AI as that force multiplier that can augment human decision-making without replacing it altogether. As these technologies mature further and commanders become increasingly familiar with them, that part of defense operations with AI is also likely to grow.
In the short term, generals and their artificially intelligent chatbots are mastering tag-teaming with each other, discerning what can be accomplished while retaining awareness of what can be amiss. The future of military command may well depend on whether human commanders can collaborate with artificial intelligence.
