Lt. Gen. Jack Shanahan, head of the Joint AI Center for the DOD, talks integrating AI into its weapons systems, building predictive maintenance tools, and more
Lt. Gen. Jack Shanahan doesn't want killer robots—but he does want artificial intelligence to occupy a central role in warfighting.
Context & Ripple Effects
The Joint AI Center's mandate has been scaling since the Defense Department committed a reported $1.7B over five years to stand it up — and this Wired interview with its founding director, Lt. Gen. Jack Shanahan, is the clearest statement yet of what that money is meant to buy: AI woven into existing weapons systems, starting with unglamorous predictive maintenance rather than autonomous targeting.
Shanahan's explicit rejection of 'killer robots' frames the center as an integration office, not an autonomy program — though the related coverage shows autonomy advancing elsewhere, with DARPA's Air Combat Evolution program pursuing AI that can fly and fight without a human pilot. The tension between those two tracks defines the story.
First-order effects
- Predictive maintenance becomes the Joint AI Center's proof-of-concept: if AI reliably cuts downtime on DOD platforms, it gives Shanahan the operational wins needed to justify expanding into weapons-system integration.
- Defense vendors selling to the DOD now face a customer whose leadership publicly insists AI be embedded across its systems, shifting procurement conversations toward data readiness and model integration.
Second-order effects
- The center's later reorganization under CTO Nand Mulchandani — described in interviews about its Silicon Valley-style structure and workflows — shows the integration push forcing the Pentagon to import commercial software practices to move at vendor speed.
- As the DOD pushes for safe adoption, its AI leadership ends up lobbying industry directly: Chief Digital and AI Officer Craig Martell's call for companies to be more transparent about their AI software makes vendor disclosure a de facto condition of defense work.
Third-order effects
- If the pattern holds, the military's AI posture splits into two lanes — human-supervised applications like maintenance moving fast under the Joint AI Center, while lethal-autonomy questions get deferred even as programs like Air Combat Evolution advance them — leaving the 'killer robots' boundary contested rather than settled.
- Sustained multi-billion-dollar investment points toward a standing defense-AI bureaucracy (the center eventually reporting up through a Chief Digital and AI Officer), institutionalizing AI oversight inside the DOD much as cyber and acquisition functions were institutionalized before it.
The trend: Military AI is being industrialized through dedicated central offices that start with logistics and maintenance use cases while deferring — but not resolving — the question of autonomous weapons.