The U.S. military has grown increasingly reliant on space infrastructure for targeting, communications, coordination, and reconnaissance over the past several decades. This reliance has led to concerns within the military about future conflict in which American access to such assets may not be assured. A National Reconnaissance Office (NRO) effort to develop automated satellite maneuvering and data analysis may provide a new approach to space that could mitigate these concerns.

The NRO wants to improve its ground systems to automatically accept data from satellites, control those satellites, and then distribute the collected information as needed. The idea is that such an artificially intelligent system could catch information of interest much quicker than a human operator—and then employ and move other space assets to investigate further if needed.

This approach could mitigate concerns about the vulnerability of American space assets and U.S. reliance on them in several ways. First, if the system proves more capable at coordinating satellite constellations and tasking satellites across missions in real-time, the approach could create a more flexible satellite network. If communications with one satellite goes down, either because of a technical failure or belligerent action, the system might be able to fill the gap with another asset faster than a human could. This rests on unproven technology, of course, but the concept is appealing.

Second, such a system would permit the deployment of more redundancies in American constellations. The NRO has been taking steps to shift towards deploying smaller, cheaper satellites. These satellites have less capability than the larger ones the NRO is known for, but are easier to deploy and replace. This means that the United States can create layers of systems that are more resilient against disruption. Dispersing capabilities into separate satellites also means that losing one only disrupts one toolset. But this sort of complex, overlapping and redundant network would be difficult to coordinate and manage without the type of automated, intelligent system the NRO is pursuing.

Finally, because the system could allow greater use of cheaper tools like cubesats—very small satellites—American space infrastructure would become cheaper overall. This would free up funds for replacements, if needed. Given the costs associated with launching large satellites, better use of smaller ones could allow faster and more flexible upgrades of in-space systems. This would spread the cost of upgrading in-orbit systems over more time, reducing exposure to large spikes in costs. This would make the entire enterprise more resilient.

All of this combined could lead to decreased risks to American interests. American reliance on space has attracted attention among would-be military rivals. Increasing resiliency may increase deterrence by removing a weakness that might tempt a near-peer power during tense international incidents. Additionally, an automated and intelligent system might be able to flag useful intelligence quicker than a human could, possibly catching something that might be overlooked and allowing much more information to be processed.

There will, however, be some issues that the NRO and U.S. national security community will have to work through. Automated maneuvering in space may become trickier as space becomes more crowded, especially as the private space industry takes off. The ability to maneuver may be restricted by other countries’ satellites as well. Much of the capability to process data is classified, but a significant part of this concept rests on the system being smart enough to properly triage incoming information. The viability of pursuing such a system is beyond an unclassified analysis, but the NRO should be wary of relying too much on unproven technology. Other national security information systems have had problems due to overly optimistic estimates.

These are big issues, but they are not insurmountable. Success will require judicious consideration before investing too deeply into the system. The NRO is deploying prototypes to test the system, which is a good thing. But prototyping is not a panacea—the NRO will have to properly apply the lessons learned. If the system does not work because of capability gaps, for example, the NRO should focus on mitigating those gaps rather than relying on unproven technology for the full roll-out stage of the program. Good parts of the system can be pursued while underperforming parts should be abandoned, allowing progress while maintaining political support.

The NRO should also look to its past. The previous system, the Mission Integration and Development Program, came in under-budget and on time. While the new system is much more complicated, the NRO should examine why the previous system succeeded in its development, and work to see if approaches and tools can be carried over.

In terms of automated maneuvering, the military will have to balance the necessary secrecy of how the system responds with the measure of transparency needed to prevent other countries from misreading moves as possible attacks on their space assets. Crowdedness in space will have to be dealt with regardless of this system, so its use may drive more coordination among spacefaring countries and companies. With proper development, however, this system could improve how the United States leverages its space assets.