Photo credit: www.technologyreview.com
At present, human operators often serve as intermediaries between systems created by various manufacturers. For instance, a soldier might need to manually adjust a camera to scan an area for potential drone threats. If they detect a threat, they must then relay that information to another soldier responsible for neutralizing the drone, typically through basic messaging applications reminiscent of early instant messaging platforms. This process is time-consuming and presents operational challenges. The Pentagon is addressing these interoperability issues through initiatives like the Joint All-Domain Command and Control (JADC2) strategy, which aims to streamline communications across various military systems.
Chris Brose, who previously served as the staff director for the Senate Armed Services Committee, highlights a long-standing awareness of interoperability issues within military systems. As Anduril’s chief strategy officer, he argues that software solutions deserve as much funding as traditional hardware like jets and aircraft carriers. In the past year, Anduril invested nearly $1.6 million in lobbying efforts and has established connections with the incoming Trump administration, given the firm’s founder, Palmer Luckey’s longstanding support for Trump, and Senator JD Vance’s investment ties with the company.
Moreover, defense hardware often struggles with connectivity challenges. Tom Keane, a senior vice president at Anduril, illustrates this issue with a relatable analogy from civilian technology: when a phone is turned off and then back on, any missed messages are still accessible. This fundamental feature is frequently absent in many defense systems, leading to the potential loss of vital data and communications in military operations. Anduril, however, claims that its systems retain data locally, mitigating this problem.
Unlocking AI’s Data Potential
The military’s push for advanced AI-integrated hardware may initiate one of the most expansive data collection efforts ever seen within the Pentagon. Companies like Anduril and Palantir are poised to play significant roles in this transformation.
On December 6, Anduril announced a partnership with Palantir, aiming to compile an extensive array of data collected through its Lattice system, including sensitive classified information essential for training AI models. This collaboration comes on the heels of Anduril’s partnership with OpenAI, which began on December 4, to enhance model-building capabilities using data from diverse military sensors. Earlier this year, Palantir proposed its AI resources to assist the Pentagon in rethinking its approach to the classification and management of sensitive data, underscoring a growing recognition of AI’s utility in military applications. As Anduril’s founder, Palmer Luckey, noted during an October interview, there has traditionally been a scarcity of robust information related to classified matters available for AI training; this collaboration may signal a shift in that paradigm.
Even though parts of this military data are already being aggregated, the transformative power of AI is set to unlock its potential. Emelia Probasco, a senior fellow at the Center for Security and Emerging Technology at Georgetown University, highlighted that the current advancements will enable the Defense Department to utilize existing data in innovative ways, potentially enhancing accuracy, precision, and speed in information processing.
The result of these advancements may concretely place AI models at the heart of military decision-making processes, rather than simply serving as tools for information extraction. This shift has drawn attention and criticism, notably when reports surfaced indicating that Israel employed advanced AI to analyze intelligence and compile target lists. Human Rights Watch pointed out that such tools often depend on inaccurate data, raising ethical concerns about their implications for civilian safety, as highlighted in a report.
Probasco anticipates a trend toward increasingly integrating AI, including generative AI, within military decision-making frameworks. She examined a system established in 2023 known as Maven Smart System, which allows military personnel to access and analyze sensor data from multiple sources, applying computer vision algorithms to assist in identifying and selecting military targets.
Source
www.technologyreview.com