Later this month in Geneva, states will gather for the Review Conference of the Convention on Certain Conventional Weapons and decide whether to begin negotiating binding rules on autonomous weapons.
More than 70 states have already indicated their readiness to do so. Meanwhile, Washington is moving in the opposite direction.
The gap between Washington and its allies on autonomous weapons has direct implications for how responsibility is built into military technology.
On 30 September, the US department of war announced an Autonomous Warfare Command, meant to scale autonomous warfare across the US military by October 2027.
The same speech launched Project Meridian, a 120-day study of the technologies the US military will need for future wars. The project will be led by tech oligarch Elon Musk, Palmer Luckey (CEO of defence firm Anduril), and US politician Newt Gingrich, with the group’s findings due in January.
Ethical and commercial issues abound. SpaceX is a backbone of US national security, and Anduril sells jammers, interceptors, and un-crewed combat aircraft to the US military.
Critics have noted that Musk and Luckey would be identifying needs that their own companies already meet, and the announcement describes no recusal or disclosure mechanism.
Others, such as professor Elke Schwarz from the. Queen Mary University of London, have long shown how the commercial logic of technology firms shapes the ethics of violent technologies, treating moral and legal questions as engineering problems to be optimised away.
Read more
Russia’s faith in ‘error-free’ killer robots and global divisions expose the moral vacuum in AI warfare
From a technical and legal perspective, this is neither possible nor desirable. Project Meridian promises that this logic will set the agenda.
What this agenda fails to cover is just as consequential. The US administration of president Donald Trump has dismantled the machinery that teaches militaries about civilian harm.
The Pentagon’s Civilian Protection Center of Excellence fell from 40 staff to nine in 2025, and the Pentagon’s own inspector general concluded the department no longer has the personnel and capabilities needed to meet its legal responsibilities.
Organisations like Airwars, which document civilian harm from military operations, find themselves offering evidence to commands that have almost no one left to receive it, let alone learn lessons to improve practices over time.
Autonomy and AI-enabled targeting depend on high quality data and feedback. A system is only as good as what its operators learn from its errors.
Research on the quantification logics of targeting, and on outdated intelligence feeding strike decisions, shows how easily speed and scale displace the judgment that precautions require.
Accelerating autonomy while abandoning the ability to learn from past practices does not produce a more precise or legitimate military.
It produces a faster one that repeats its mistakes, which will continue to exacerbate civilian harm in new theatres of war.
Read more
First fully-autonomous drone kill Ukraine heralds ‘dystopian future of untethered weaponry’
Washington exporting normalisation
This issue is relevant to discussions in Geneva because technology that is built, bought and normalised in Washington is sold elsewhere.
European militaries integrate, procure and operate alongside American systems.
And the diplomatic track is losing ground on exactly this point: the Group of Governmental Experts’ final report softened what earlier drafts framed as obligations on responsibility and accountability into encouragement.
This moment also offers opportunities. Europe is in a position to answer with more than just words.
Responsibility works best when it is designed into a system from the start.
Read more
Why Pope Leo is right to call on EU to regulate lethal AI weapons
Precautions must be built in at the design stage, with deliberate friction where speed would erode judgment, meaningful human decision-making, and sustained engagement among developers, lawyers, and operators.
States already have duties to review new weapons and to protect civilians. Applying these across the whole lifecycle of a system takes institutional will, and Europe is uniquely positioned to demonstrate this.
This will also strengthen European sovereignty. Relying on a handful of American vendors means importing their design assumptions along with their products.
European standards for testing and assurance, limiting civilian harm written into procurement contracts, and legal review that begins at the design stage would give governments real choices about what they buy and on what terms.
Any supplier, American or European, would have to meet them.
In Geneva, European governments should back the start of negotiations on a legal instrument, and they should write responsibility by design into their own procurement now, without waiting for a treaty.
Europe should demand a different theory of responsibility from whoever supplies its autonomy.