In Tampa, Florida, U.S. special operations forces are seeking advanced AI solutions that can harness the capabilities of large data centers in remote, disconnected environments.
Rob McClintock, the program manager for intelligence at the program executive office for digital applications, reported that Special Operations Forces (SOF) are already leveraging generative AI for essential tasks like resource management and troop deployment. They are also exploring its potential for tactical operations. However, existing AI tools predominantly rely on cloud infrastructure linked to extensive data centers, making them impractical for use in isolated locations. By bringing computing capabilities closer to the frontline—the so-called "tactical edge"—operators can access critical data more swiftly and enhance their decision-making processes, as discussed during the SOF Week event organized by the Global SOF Foundation.
To address these challenges, Special Operations Command is in search of systems that can deliver cloud-like processing power directly at the sites where data is gathered and utilized. This is part of a broader development trend known as "fog computing." Additionally, SOCOM is interested in more efficient large language models that maintain a deep understanding of human instructions while requiring less computational resources.
Colonel Robert "Ramsey" Oliver, PEO of SOCOM’s SOF Warrior program, highlighted the potential of voice commands to ease the cognitive burden on operators. Current AI and tech solutions from leading consumer companies often do not cater to the specialized needs of tactical operations. According to Melissa Johnson, SOCOM’s acquisition executive, innovative solutions are likely to stem from agile startups rather than the larger tech firms with rigid approaches. She emphasized the need for a diverse acquisition strategy that includes dynamic AI solutions from smaller enterprises.
The primary goal for SOCOM in adopting AI tools is to streamline operations, simplifying processes like coordinating various drone systems or conducting missions based on minimal verbal or physical cues, as noted by Lt. Col. Aaron Davidson, who oversees the unmanned systems autonomy and interoperability portfolio. Furthermore, McClintock mentioned that they are investigating the potential for AI "agents" capable of strategic planning, revisions, and execution of operations.
