The US Army is building its own Chatbot for combat

the US Army It is developing artificial intelligence models trained on data from real missions, with the aim of deploying a chatbot specifically for soldiers.

“We learned all these lessons from missions like the Russia-Ukrainian war and Operation Epic Fury,” Alex Miller, the Army’s chief technology officer, says in an interview with WIRED. “There is a tremendous amount of knowledge available.”

Miller showed WIRED magazine a prototype of the system, called Victor, which combines a Reddit-like forum with a chatbot called VictorBot to help troops display useful information, such as how to best configure electromagnetic warfare systems for a given mission. When a Soldier asks how to set up their device, VictorBot generates an answer and points to relevant posts and comments from other service members. “Electromagnetic warfare is a very difficult topic,” Miller says. He adds that Victor “can generate a response and cite all the lessons learned from the (different) modules.”

The Pentagon has ramped up its efforts to integrate AI into military systems over the past two years, but Victor is a rare example of the military building AI for itself. The project demonstrates how keen the US Army is to master the basics of artificial intelligence, and how this technology can change the daily lives of many troops.

Miller says the Army is working with a third-party vendor that will operate and fine-tune the AI ​​models powered by Vector. He declined to name the specific company because the contract has not yet been announced. He says more than 500 data repositories have been introduced into the system, and notes that Vector will seek to reduce the potential for errors in a similar way to business chatbots, by citing real-world sources.

Efforts to integrate AI into military systems accelerated following the introduction of ChatGPT in 2022. More recently, AnthropicTechnology reportedly played a prominent role A role in planning operations in Iran Through a system powered by Palantir.

As these systems become more capable, disagreements have emerged over how to deploy AI. Earlier this year, Anthropics came face to face with the PentagonArguing that its technology should not be used to operate autonomous weapons or monitor American citizens.

Same errors

Vector is being developed within the Combined Arms Command (CAC). It’s not uncommon for different brigades to make the same mistakes on different missions, says Lt. Col. John Nielsen, who oversees CAC’s work in Victor. He adds that the goal of Vector is to eventually make the system multimedia so soldiers can feed in images or video and get insights. “Victor will be one of the only sources with access to official Army information,” Nielsen says.

Project Vector highlights the potential for AI to automate a lot of unsexy back-office tasks within the Defense Department, says Lauren Kahn, a senior research analyst at the Georgetown Center for Security and Emerging Technology and a former policy adviser to the Pentagon. Late last year, the department introduced GenAI.mil, an initiative aimed at spurring greater adoption of AI among DoD employees.

However, if Victor proves successful, Khan believes the Army could eventually hire a large AI company to enhance the system’s capabilities. “It’s clear that large labs will have a comparative advantage” in terms of building and deploying the latest AI technologies, she says.

Intel malfunctions

Artificial intelligence could present new kinds of problems for militaries, says Paul Sharry, CEO of the Center for a New American Security and a former U.S. Army Ranger. AI models’ tendency to flatter can be particularly problematic, says Shari. “I can imagine situations where this might be particularly troubling in the context of intelligence analysis,” he explains.

Shari adds that AI adoption may become more complex as systems advance from chatbots to agents capable of using software and computer networks. “Agent AI raises this whole new set of security-related challenges,” he notes.