The U.S. military must develop reliable military AI, including autonomous weapons, and adapt its doctrine and warfighters, or fall behind
Source: https://www.facebook.com/americanspectator/. "Artificial Intelligence and the Future of Warfare: Can the U.S. Military Keep Up? | The American Spectator | USA News and Politics." November 1, 2026. spectator.org
The Gist
The author says AI weapons are already being used by Russia, Ukraine, and probably China. Future threats like drone swarms and hypersonic missiles will move too fast for humans to approve each shot. So the U.S. has to build reliable AI, give it some autonomy, and train commanders to work with it as a trusted partner. If it doesn't, it will fall behind or lose control of decisions to machines.
Conclusion
The U.S. military must develop reliable AI systems, accept some autonomous lethal decision-making, and build the doctrine, trust-based human-AI teaming, and cognitive skills to use them. Otherwise it will fall behind adversaries and leave decisions entirely to machines.
Premises
- Russia and Ukraine already use AI to make kill/no-kill decisions in combat, and China is reportedly developing similar systems.
- Past arms races (nuclear weapons, missiles after Sputnik) show that falling behind a rival in a decisive military technology is dangerous, so a military AI gap is a real danger if the U.S. does not keep developing reliable AI.
- Emerging threats such as drone swarms and hypersonic missiles mixed with conventional weapons will arrive faster than human decision-making can handle, so commanders will have to grant AI weapons autonomous release authority.
- In the author's simulations, human teams and an AI system made wrong engagement decisions at the same rate (20%), though the AI decided faster. Human judgment is therefore not clearly more reliable than AI's, which weakens the case for keeping humans in every loop.
- Democracies have an advantage in human-AI teaming because their militaries are built on earned, two-way trust rather than fear. Treating AI as 'slave soldiers' would backfire, while treating it as a partner could produce a true partnership.
- Current U.S. doctrine and acquisition are lagging. The Marine Corps' bet on the short-range, subsonic, hard-to-relocate NMESIS system, which hypersonic missiles outmatch, shows senior leaders have not grasped AI's implications for shooting and moving.
- The only alternative to smarter, more agile warfighters who can work with AI is to leave decision-making entirely to machines, which should frighten everyone.
Assumptions
- Adversary use of AI weapons means the U.S. must match them rather than seek restraint or arms control.
- The speed advantage of autonomous systems outweighs the moral and legal costs of added civilian casualties.
- A single simulation with a small sample generalizes to real combat decisions.
- AI systems can meaningfully respond to 'trust' and respectful treatment, or the analogy to human soldiers' morale is apt.
- Democratic and autocratic militaries differ systematically in how they build trust, and this difference carries over to AI integration.
- Human-in-the-loop control is infeasible against the fastest threats, and no alternative safeguards are available.
- Reliable AI is achievable and the U.S. can develop it fast enough to matter.