AI's Autonomy: Navigating Ethical Choices
Explores ethical puzzles in AI-driven autonomy, like self-driving cars facing unavoidable accidents or autonomous weapons making life-or-death decisions without human intervention.
Topics: technology, politics, individual and values
Example
Lethal autonomous weapons systems can select and attack targets without a human deciding each strike. On 22 December 2023, the UN General Assembly passed its first resolution on them, by 152 votes to 4 with 11 abstentions, stressing the 'urgent need' to address the risks and asking the Secretary-General for countries' views. A follow-up resolution in December 2025 passed 164 to 6. The core ethical question is who is morally responsible when a machine decides to kill, and whether a human must always keep control.
Key facts
- On 22 December 2023 the UN General Assembly passed its first resolution on lethal autonomous weapons by 152 votes to 4, with 11 abstentions.
- A further General Assembly resolution on autonomous weapons passed in December 2025.
Limitations
- General Assembly resolutions are not binding, and no treaty bans or regulates these weapons; the ethical debate is ahead of the law.
Evaluations
technology evaluation
Support
Machines can react faster than people and do not panic, so supporters argue autonomy could reduce mistakes in fast-moving combat.
Counterargument
AI systems can misidentify targets in ways that are hard to predict or explain, and a machine cannot judge surrender or intent as a soldier can.
Rebuttal
The more a system acts on its own, the more its errors become no one's decision, which is why many states call for meaningful human control.
Additional support
Drones with growing autonomy have been used in recent wars, so the question is no longer theoretical.
politics evaluation
Support
A 152 to 4 vote shows most governments see a need for international rules on autonomous weapons.
Counterargument
Major military powers prefer national guidelines to binding limits, and talks in the Convention on Certain Conventional Weapons need consensus, which lets any state block a treaty.
Rebuttal
Moving talks to the General Assembly, where majorities decide, is one way supporters are trying to get around that veto.
Additional support
The arms race logic is strong: if rivals build autonomous weapons, each state fears falling behind, the classic barrier to arms control.
individual and values evaluation
Support
Many ethicists argue that a decision to take a human life should never be handed to a machine, because it denies the victim's dignity.
Counterargument
Others reply that what matters morally is fewer deaths, and if machines kill fewer civilians than tired soldiers, that is better.
Rebuttal
Even if machines were more accurate, responsibility would still blur, and law and ethics depend on someone being answerable.
Additional support
Tech workers at several firms have refused to work on weapons projects, showing individuals acting on these values.
Sources
- United Nations Digital Library (2023-12-22): Lethal autonomous weapons systems: resolution adopted by the General Assembly (A/RES/78/241)
- Human Rights Watch (2024-01-03): Killer Robots: UN Vote Should Spur Action on Treaty
- United Nations Digital Library (2025-12-01): Lethal autonomous weapons systems: resolution adopted by the General Assembly (A/RES/80/57)