The United States Department of Defense (DoD) has initiated tests to evaluate the effectiveness of generative AI in assisting with military planning, including responses to potential global conflicts and streamlining internal information access. Collaborating with undisclosed allies, the DoD’s digital and AI office at the Pentagon is conducting experiments involving five large language models (LLMs), one of which is Scale AI’s “Donovan” model.
Colonel Matthew Strohmeyer of the Air Force acknowledged the success of initial LLM tests, describing them as “highly successful” and impressively fast. However, he admitted that the technology is currently not ready for full implementation. One test demonstrated the LLM’s capability to deliver requested information within a mere 10-minute timeframe, a significant improvement compared to the usual days-long process involving multiple personnel.
These LLMs have been provided with classified operational information to generate responses related to real-world situations. The ongoing tests aim to determine whether they can assist in planning responses to potential escalations, such as the tense military situation with China in the Indo-Pacific region.
While the current testing period is set to conclude on July 26, the exploration of AI’s potential in warfare has been underway for some time. In May, the British government’s Defence Science and Technology Laboratory (Dstl) hosted the first joint trial involving the US, Australia, and the UK, where AI-enabled military drones were tested for target tracking and detection. Notable achievements during this trial included retraining AI models in real-time while in flight and interchanging the models between the AUKUS alliance, with the aim of swiftly integrating these technologies into military capabilities.
This development signifies an important step in assessing the viability of AI in military operations, particularly in complex global conflicts. While challenges and limitations remain, the successful implementation of AI has the potential to enhance response times, information gathering, and decision-making processes within the defense sector.
The ongoing experiments conducted by the US Pentagon highlight the growing significance of AI advancements in bolstering military capabilities and planning strategies in an increasingly complex global landscape.