Cheap Autonomy Breaks the Chain of Command

Cheap autonomy has already arrived on the battlefield, and the first institution it breaks is not the army but the chain of command. The question is no longer whether autonomous weapons will spread but who, if anyone, remains in the decision loop when one fires.

Yaroslav Azhnyuk, who builds autonomous drone software from Kyiv after fourteen years on pet cameras, frames the moment inverting the usual ethics debate. On Craig Smith's Eye on AI podcast, he argues that within five to ten years using weapons without AI may itself be considered unethical, because non-autonomous systems cause more collateral damage. The provocation is real; it is also the smaller point.

The risk is not Skynet but the subtle transfer of power when a hundred smarter advisors stop waiting for the President to decide, and nobody notices the handoff. Azhnyuk's own number sets the scale: a $500 drone that can already destroy a $5 million tank becomes roughly ten thousand times more capable when full autonomy is added for a few hundred dollars more. The economic engine is settled. The governance engine is not.

The falsifier is plain. Ask a defense minister, a general, or your own government whether the lethal decision can be reconstructed after the fact in plain language: auditable logs, a named human in the loop, an escalation rule. If the answer is no, the software stack has already outranked the chain of command.