autonomous weapon systems
Weapons that can select and engage targets without direct human control, raising distinct legal and responsibility questions.
As defined in 46 | Kate Darling on Our Connections with Robots
Where it comes up · 7
- Apocalypse Now? Peter Thiel on Ancient Prophecies and Modern Tech3 claims
- 138 - The Edge Of Humanity2 claims
- 30 | Derek Leben on Ethics for Robots and Artificial Intelligences2 claims
- 66 - Living With Robots1 claim
- Yuval Noah Harari1 claim
- 46 | Kate Darling on Our Connections with Robots1 claim
- TimesTalks | Yuval Noah Harari1 claim
Most important claims · 10 of 11Search all claims for “autonomous weapon systems” →
- AI is the first tool in history that can make decisions by itself; all previous tools (a knife, an atom bomb) empowered humans because they could not decide anything—the atom bomb cannot choose which city to bomb—whereas autonomous weapon systems can decide by themselves who to kill, which is unprecedented.Yuval Noah Harari
- Humanity faces three problems—nuclear war, climate change, and technological disruption—that by their nature can only be solved through global cooperation: no country is ecologically independent so you cannot build a wall against rising temperatures, and AI and biotech are global so a national ban on autonomous weapons or human genetic engineering merely lets rivals proceed and triggers an arms race; the only way to prevent such an arms race is global cooperation.138 - The Edge Of Humanity
- Global cooperation does not require a single world emperor but real trust between countries, as exists in the EU; AI is harder to govern than nuclear weapons because nuclear development cannot be hidden but AI can be developed in secret, making verification and trust the key obstacle—yet the France-Germany reconciliation shows that even bitter enemies can build durable, credible trust through hard work, so China-US trust is not hopeless.138 - The Edge Of Humanity
- There are multiple categories of existential risk including nuclear war, bioweapons or dangerously engineered bioweapons, AI and autonomous weapon systems guided by AI, and environmental risks beyond climate change that require caution about planetary harm.Apocalypse Now? Peter Thiel on Ancient Prophecies and Modern Tech
- Autonomous weapon systems (killer robots) cannot be effectively regulated through unilateral national bans or even signed international agreements without enforcement mechanisms, because unlike nuclear weapons, AI development is much easier to hide and would require unprecedented levels of trust between nations like the US and China.TimesTalks | Yuval Noah Harari
- Autonomous weapon systems that can decide to kill without a direct human command undermine accountability because international law requires showing intent for war crimes, and intent is far harder to establish when a machine rather than a person makes the lethal decision—removing responsibility for harm in a way critics argue is unethical and possibly illegal.66 - Living With Robots
- Autonomous weapon systems have internal logic driving their development: human operators in the loop get jammed, so there's systemic pressure to make them fully autonomous, creating a first-strike weapons dynamic that intersects with US-China arms race dynamics in unpredictable waysApocalypse Now? Peter Thiel on Ancient Prophecies and Modern Tech
- The problem of assigning responsibility when autonomous robots cause harm is not as new as people think; since the earliest laws humans have grappled with liability for autonomous agents like an ox that gores someone in the street, and these historical (culturally varied) solutions—even a 17th-century 'law of the horse'—can be drawn on to regulate autonomous vehicles and weapons.46 | Kate Darling on Our Connections with Robots
- Totalitarian one-world government represents an existential risk as significant as nuclear war, bioweapons, and weaponized AI, because once established it cannot be reversed, and it appears to be the implicit solution offered by advocates of managing other existential risks through centralized control.Apocalypse Now? Peter Thiel on Ancient Prophecies and Modern Tech
- For security and military robots, contractarianism via maximin would favor neutralizing a threat over killing the threatening agent, because killing makes that agent maximally worse off while neutralization is the best of possible outcomes; an ideal war robot ordered to kill could refuse and instead apprehend, analogous to military ethics rules against obeying orders to commit war crimes.30 | Derek Leben on Ethics for Robots and Artificial Intelligences