The U.S. Air Force's experiment with a ChatGPT-like AI tool has garnered significant attention across its ranks. Known as NIPR GPT, the system was introduced in June 2024 and has already attracted more than 80,000 users within the Air Force and Space Force. This generative AI tool, designed to enhance productivity and streamline tasks, has been embraced by service members for a wide range of applications, from content creation to coding, and much more.
Developed by the Air Force Research Laboratory (AFRL), the Non-classified Internet Protocol Generative Pre-training Transformer (NIPR GPT) was created to assess how AI could boost efficiency and effectiveness within the armed services. Early users of the AI come from diverse career fields, and the Air Force is using their feedback to improve the tool and determine its long-term potential. Alexis Bonnell, the AFRL’s chief information officer, highlighted that the tool's widespread adoption demonstrates a strong demand for AI-driven solutions to support daily operations.
New from 404 Media: the Air Force is researching a 'https://t.co/QSjV9Uno6v' that would adapt drones to new environments. Contract includes "expand data lake for extensive communications data." Documents here: https://t.co/TicNU1APyF pic.twitter.com/CDAPSPJx0J
— Joseph Cox (@josephfcox) September 16, 2024
NIPR GPT allows personnel to handle time-consuming tasks like file searches and routine queries with greater speed, allowing airmen and guardians to focus on mission-critical duties. In particular, the AI’s use of "retrieval-augmented generation" (RAG) has been cited as a breakthrough, enabling users to tailor the tool to meet specific needs.
This entrepreneurial approach, as Bonnell described, gives personnel more control and agility, eliminating the need for long wait times for software updates or system improvements.
Testing ChatGPT Advanced Voice Mode’s ability to create sounds.
It somewhat successfully sounds like an airline pilot on the intercom but, if pushed too far with the noise-making, it triggers refusals. pic.twitter.com/361k9Nwn5Z— Cristiano Giardina (@CrisGiardina) July 31, 2024
The Air Force is not alone in exploring AI applications. Both the Navy and Army are actively testing similar systems. The Navy introduced its AI, Amelia, in 2023, while the Army has integrated the Ask Sage AI platform into its operations. However, the Air Force’s NIPR GPT pilot stands out for the speed at which it has been embraced across multiple service branches.
One of the key aspects of NIPR GPT’s success is its flexibility in acquisition strategies. The Air Force has yet to commit to a specific vendor, opting to rely on open-source AI models. This allows them to keep the system adaptable and test new commercial tools as they become available. As Bonnell emphasized, the real-world data being collected from the pilot will be invaluable for future AI investments, allowing the military to make evidence-based decisions in negotiations with tech vendors.
This rapid growth in AI usage signals a shift in how military operations will be conducted in the future. While NIPR GPT is currently designed for support roles, the AFRL is continuing to explore how AI could be integrated more broadly, potentially even into combat scenarios. Although the focus remains on non-classified applications for now, the success of the NIPR GPT pilot suggests a future where AI systems play a central role in both combat and logistical operations across the U.S. military.
As the AI continues to evolve, Bonnell noted that ongoing assessments will focus on identifying what tools are most valuable to service members. The Air Force intends to refine the system based on user feedback, ensuring that NIPR GPT remains a crucial asset for the military's digital transformation.
