The Department of Defense, first under the Biden administration, but continuing under the Trump administration, is having far too much trouble deploying “AI weapons,” even figuring out how to use those it has deployed. An illustration of this is this:

The Pentagon has also struggled to find software that can successfully control large numbers of drones, made by different companies, working in coordination to find and potentially strike a target—a key to making the Replicator vision work.

This isn’t so much a software commonality or interoperability problem as it is a problem with bureaucrats giving too much weight to the desires of contractors and their lobbyists, each of whom insist on their own proprietary software. What’s needed here isn’t more money (although sometimes that helps), it’s an operational officer, one fresh in from a combatant command that’s facing current conflicts, being put in charge of AI development and programs. This officer, independent of the lobbyists, needs to write a better requirements document, one that specifies the standards of commonality (not merely interoperability—this is software, not hardware) that each contractor and contractor wannabe must meet in order for their AI package to be considered.

Beyond that, this AI OIC must open his Requests for Proposals far beyond the major contractors and take in RFPs from small businesses and startups—that’s where the ideas and agility exist—then push the development and acquisition pace, eliminating the layers of bureaucrats’ reviews.

Finally, this AI OIC must be willing to spend money on mistakes and dead ends rather than demanding (near) perfection at one stage before moving on to the next.

That’s just the development side. The AI OIC also must have four or five teams of operational officers—again, fresh in from combatant commands—working on tactics and strategies that make use of the four or five leading AI weapons candidates to investigate best uses for integrating these weapons into a combatant command unit’s (at all levels) existing suite of weapons. That tactics and strategies development effort should also lead to adjustments in those units’ existing tactics and those commands’ existing strategies, even their existing suites of weapons.

There’re just too much bureaucracy and too little operational consideration in the currently extant DoD. SecDef is making progress on the bureaucracy qua bureaucracy, but he’s got a long way to go. It’s time for him to zero in on lethality.