UN and Red Cross demand global rules for AI weapons as fears grow over machines targeting humans
António Guterres and ICRC President Mirjana Spoljaric are demanding legally binding rules for autonomous weapons as military technology advances faster than international regulation.
Follow Nukunya
Get the latest breaking news and updates from Nukunya.
António Guterres and ICRC President Mirjana Spoljaric say the world is nearing a “moral red line” as weapons become more autonomous. Their call for binding international rules has particular consequences for Africa, where drone warfare, insurgencies and externally supplied military technology are already reshaping conflict.
GENEVA – The United Nations and the International Committee of the Red Cross have renewed an urgent call for legally binding international rules on autonomous weapons, warning that advances in military technology are bringing the world dangerously close to allowing machines to select and attack human beings with diminishing human control.
UN Secretary-General António Guterres and ICRC President Mirjana Spoljaric said states should move beyond years of discussion and begin negotiating clear prohibitions and restrictions on autonomous weapon systems. Their joint appeal, issued on 25 August, warned that weapons technology has advanced faster than the rules intended to govern it.
“We are now dangerously close to crossing a moral red line,” they said, identifying that line as the autonomous targeting of humans by machines. The warning is significant partly because the technology is no longer treated by international organisations as a distant theoretical concern.
Autonomous functions already exist in some weapon systems, and militaries are increasingly experimenting with artificial intelligence, drones, sensors and automated decision-making.
But an important evidentiary boundary remains.
UN and ICRC officials said on 25 August that there is no confirmed evidence that fully autonomous weapons have yet been used to target human beings. Reports that fully autonomous or AI-guided drones are already doing so remain unconfirmed.
That distinction matters.
The international argument is increasingly about setting limits before technology reaches the point where machines routinely make lethal targeting decisions without meaningful human judgement.
THE FIRST DISTINCTION
What is an autonomous weapon?
An autonomous weapon is not simply a drone, and a weapon does not necessarily need artificial intelligence to be autonomous.
ICRC DEFINITION
Once activated, an autonomous weapon system can select and apply
force to targets without further human intervention.
AUTONOMY
Who or what selects and engages the target?
AI
A technology that may be incorporated into a system.
Human decides
A human operator selects the target and decides when force should be used.
System performs the selection
Once activated, programmed criteria can determine how the weapon identifies and applies force to targets.
THE CENTRAL QUESTION
The most important legal and ethical question is not whether software is present.
Who retains meaningful control over the decision to use lethal force?
THE PROPOSED RULES
What do the UN and Red Cross want governments to do?
Their appeal is not a proposal to prohibit every autonomous function in every weapon.
01
PROHIBIT
The most problematic systems
- Weapons whose effects cannot be sufficiently understood or predicted.
- Autonomous weapons designed or used to target human beings.
02
STRICTLY RESTRICT
Other autonomous weapon systems
Controls would be intended to preserve human judgement and ensure compliance with international humanitarian law.
THE OBJECTIVE
A legally binding international instrument containing clear
prohibitions and restrictions.
WHY THE WARNING IS COMING NOW
Technology is advancing faster than the rules
States have spent years discussing lethal autonomous weapons under the Convention on Certain Conventional Weapons.
The Group of Governmental Experts continued work on possible elements of an international instrument.
States are due to continue developing possible elements of rules.
The ICRC regards the meeting as a major opportunity for governments to decide whether formal negotiations should begin.
THE POLITICAL CHOICE
Establish limits while governments still have meaningful choices, or attempt to regulate the technology after militaries have already integrated it extensively.
EVIDENCE CHECK
Are autonomous weapons already killing people?
This is where careful reporting matters most.
✓
SUPPORTED BY THE AVAILABLE EVIDENCE
What can be said
- Autonomous functions have existed in weapon systems for years.
- Defensive systems can react rapidly to incoming threats.
- Autonomous capabilities are increasingly being combined with surveillance, recognition and navigation technologies.
- The ICRC says existing autonomous systems have caused combatant casualties.
?
NOT CONFIRMED
What should not be overstated
- There is no confirmed evidence cited here of fully autonomous weapons independently selecting human beings for lethal attack.
- Reports concerning fully autonomous or AI-guided drones in current conflicts remain unverified.
THE REGULATORY ARGUMENT IS PREVENTIVE
Governments do not need to wait until autonomous human-targeting
weapons become routine before agreeing what should be unacceptable.
THE AFRICAN DIMENSION
Why Africa should care
Autonomous weapons are not an abstract debate relevant only to the world’s largest military powers.
AFRICAN UNION POSITION Meaningful human control
SUPPORT 40+ AU member states
have supported calls for a legally binding agreement, according to
the AU material cited in the report.
THE STRATEGIC QUESTION
Africa risks becoming a rule-taker
Many of the most advanced autonomous military technologies are being developed by states and defence industries with large research budgets, computing infrastructure and established weapons sectors.
RULE-TAKER
Adapt later to standards largely designed around the assumptions of technologically advanced militaries.
RULE-MAKER
Participate now in defining acceptable limits, accountability and requirements for human control.
THE ACCOUNTABILITY PROBLEM
If a machine makes the wrong targeting decision, who is responsible?
Legal responsibility does not disappear because software contributed to an attack.
Who authorised deployment?
Who purchased and operated the system?
Who designed the targeting technology?
Who exported or transferred it?
THE PROBLEM
The greater the distance between human judgement and the final use of force, the harder investigations and accountability can become.
For states with limited technical and forensic capacity, investigations may increasingly require expertise in algorithms, sensors, software behaviour and automated classification as well as conventional weapons law.
REAL-WORLD AMBIGUITY
A machine can function as designed and still produce an unacceptable result
Military environments do not fit neatly into predetermined categories.
Combatant, civilian, journalist, police officer or humanitarian worker?
Military transport, civilian transport or an object whose use has changed?
Military facility, home, hospital or structure serving multiple functions?
Still fighting, wounded, surrendering or otherwise no longer a lawful target?
THE ETHICAL RISK
The danger is not limited to machines malfunctioning. A system can
operate exactly as programmed while the real world remains more complex
than its categories.
GHANA’S POSITION
What Ghana’s position suggests
Ghana has acknowledged the broader challenge presented by emerging military technologies.
GHANA AT THE UN FIRST COMMITTEE, 2025
Ghana argued that existing norms should be strengthened and legal gaps addressed as autonomous weapon systems, armed drones and other emerging military technologies proliferate.
Rules governing autonomous weapons therefore matter not only to countries that manufacture them, but also to personnel operating in environments where other actors may deploy them.
THE PRINCIPLE AT THE CENTRE
What does “meaningful human control” actually mean?
A human appearing somewhere in the targeting process is not necessarily enough.
Can the human understand what the system is doing?
Can the person realistically monitor its operation?
Can the operator stop or redirect the system?
Does meaningful human judgement still determine whether lethal force is used?
WHERE THE DIPLOMACY STANDS
International concern is broad, but there is still no treaty
UN GENERAL ASSEMBLY • DECEMBER 2025 164 in favour
AGAINST 6 votes
ABSTENTIONS 7 votes
General Assembly resolutions do not themselves create the legally binding instrument being sought by the UN Secretary-General and ICRC.
NUKUNYA ANALYSIS
Africa needs a seat at the table before the rules harden
The most important African question is not whether the continent is about to build armies of “killer robots”.
That framing is too simplistic.
THE MORE IMMEDIATE QUESTION
Will African states have enough influence over the international rules
governing technologies that may eventually appear in their security environments?
A weapon does not need to be developed in Africa to affect Africans.
Rules built mainly around technologically advanced militaries may not reflect conflicts characterised by displacement, irregular armed groups, weak identification systems and combat close to civilian populations.
THE WINDOW FOR ACTION Now → Next → Later
Capability is developing.
Governments decide whether to negotiate restrictions.
Once systems become strategically embedded, restrictions become politically harder.
NUKUNYA VIEW
The UN and Red Cross warning is fundamentally about timing.
The world has not yet reached the scenario they fear. The question is whether governments establish limits while meaningful choices remain available.For African states, that means participating now rather than waiting for technologies developed elsewhere to determine the terms of the debate.
EDITORIAL EVIDENCE LOCK
What is confirmed and what is not
The UN Secretary-General and ICRC President are calling for urgent negotiations towards a legally binding instrument containing prohibitions and restrictions on autonomous weapons.
States are discussing possible elements of international rules under the Convention on Certain Conventional Weapons.
UN and ICRC officials say there is currently no confirmed evidence that fully autonomous weapons have independently targeted human beings.
Reports concerning fully autonomous or AI-guided drones in current conflicts remain unverified.
QUICK EXPLAINER
Five questions readers may still have
01 Is every military drone an autonomous weapon?
No. A remotely piloted drone can still depend on a human operator to select the target and decide whether and when to strike.
02 Does autonomous automatically mean AI-powered?
No. Artificial intelligence can be integrated into autonomous systems, but autonomy and AI describe different characteristics.
03 Is the UN proposing to ban every autonomous weapon?
No. The framework described in the source material distinguishes between systems that should be prohibited and other systems that should face strict restrictions.
04 Are fully autonomous weapons already targeting people?
UN and ICRC officials cited in this report say they currently have no confirmed evidence of fully autonomous weapons independently selecting human beings for attack.
05 Why is Africa involved in this debate?
African countries can encounter the technology through imports, partnerships, peacekeeping and conflicts involving state and non-state actors even if the systems were designed elsewhere.
HOW NUKUNYA VERIFIED THIS REPORT
Source integrity
This report is anchored primarily in institutional and primary sources rather than secondary commentary.
Primary evidence for the proposed international rules.
Clarification of what is and is not confirmed about battlefield use.
Evidence on drones, non-state actors and meaningful human control.
Verification of the 2026 expert discussions and meeting schedule.
EDITORIAL STANDARD
Claims about current battlefield deployment are kept separate from verified institutional positions. Unconfirmed reports are not converted into established facts.
Related posts:
Pope Leo XIV Urges AI ‘Disarmament’ in Major Vatican Warning on Technology Risks 
MTN Group CEO Ralph Mupita Joins UN AI Commission in Major Global AI Leadership Move 
Humanoid Robots and the Future of War: Revolution or Hype? 
Nearpays Wins UN AI for Good Innovation Factory, Becoming First African Startup to Claim Global Title 
National ICT Week 2026 
