I am far less concerned about sentient AI deciding to destroy humankind and more worried that fellow humans will deliberately direct smart AI systems to target, control, or obliterate specific cultures, countries, societies, freedoms, or political adversaries.

AI autonomous weapons, both kinetic and cyber variants, are a reality.  The most severe risks are when such tools are used to target non-combatants, ideologies, and the critical infrastructures they rely upon.

International policymakers must define and coalesce around a set of acceptable boundaries that enable deterrence and active response in strength against aggressors who intentionally direct such smart weapons against civilian and humanitarian protected targets. 

It must be done now, before desperate and authoritarian governments are lured by the strategic and coercive power of AI autonomous weapons and their supporting technologies, and begin the planning, development, and deployment of capabilities at a massive scale and impact.

Full article on Help Net Security: https://www.helpnetsecurity.com/2026/09/17/autonomous-ai-weapons-future/