Autonomous weapons put responsibility and accountability back on the table

A forthcoming UN review conference will test whether states are ready to negotiate binding rules on autonomous weapons, while the United States is expanding its military work on autonomy. In an EUobserver opinion piece, international law scholar Jessica Dorsey argues that Europe should pair treaty talks with safeguards in its own procurement, rather than wait for Washington to change course.

Watch Desk analysis

What happened

Dorsey writes that more than 70 states have indicated they are ready to negotiate binding rules at the Review Conference of the Convention on Certain Conventional Weapons in Geneva, later this month. She contrasts that push with a US announcement on 30 September of an Autonomous Warfare Command, intended to scale autonomous warfare across the military by October 2027.

The same announcement launched Project Meridian, a 120-day study of future military technology led by Elon Musk, Anduril chief executive Palmer Luckey and former politician Newt Gingrich, with findings due in January, Dorsey says. She argues that the leaders’ commercial interests make disclosure and recusal important questions; the article says the announcement described no such mechanisms.

Why it matters

Dorsey’s central concern is not simply whether a weapon uses AI, but whether institutions can assign responsibility, make meaningful human decisions and learn from errors. She points to reduced staffing at the Pentagon’s Civilian Protection Center of Excellence and argues that weak feedback and oversight could make automated targeting faster without making it more accountable. These are the author’s arguments and reported details, not a settled assessment of the systems’ effects.

European governments also buy and operate alongside US military systems. Procurement rules could therefore shape which safeguards suppliers build in, even before any international agreement is reached.

Our read

This is a serious argument about governance as much as technology. A treaty matters, but so do the requirements attached to the next contract: responsibility is easier to build in before a system is fielded than to bolt on after something goes wrong. Dorsey’s piece is an argument for action, not a neutral account of every state’s position.

What to watch

• Whether Geneva states begin negotiations on a binding legal instrument.
• What Project Meridian recommends when its findings are due in January.
• Whether European governments put human decision-making, testing and civilian-protection requirements into procurement.
• Whether future rules address the full lifecycle of autonomous systems, not just their use in the field.

Discussion spark: Should governments negotiate binding international rules for autonomous weapons first, or can procurement standards at home do more to make military AI accountable?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.