Ban on Trans Athletes May Lead to Invasive Gender Checks

Killer Robot Talks in Geneva End in Stalemate as Nations Agree Only to Keep Talking

GENEVA — Diplomats gathered in Geneva this week with an ambitious goal: to begin writing international rules for the use of autonomous weapons, sometimes called “killer robots.” When the sessions ended, representatives had agreed only to “start discussing options” after a largely fruitless week of negotiations. For humanitarian groups, that was a deeply unsatisfying result.

The talks took place under the United Nations Convention on Certain Conventional Weapons, a forum that has existed for decades as a way for states to address the most dangerous technologies used in armed conflict. For more than ten years, countries have been debating how to handle a new generation of weapons systems that can identify, track, and engage targets without direct human control. Some governments and arms-control advocates have insisted that the only safe path is a legally binding treaty that forbids the most dangerous forms of autonomy. Others, including some of the world’s most powerful militaries, have argued that voluntary guidelines and further study are better.

Autonomous weapons are broadly understood as systems that can select and engage targets with little or no human intervention. In their fully autonomous form, they would be able to decide who to kill without a human pulling the trigger. This raises profound moral and legal questions. Machines do not understand the value of human life. They cannot reliably distinguish between a person surrendering and a person reaching for a weapon. They cannot weigh the psychological pressure of a battlefield. The idea that such systems could comply with international humanitarian law is highly doubtful.

Even the term “killer robots” can obscure as much as it reveals. It conjures images of skeletally armed androids moving through ruined cities. In reality, the systems under discussion are often unremarkable pieces of machinery: drones, sensors, targeting software, and automated gun systems. But the capabilities are no less consequential. A small autonomous drone with facial recognition could, in theory, be used to carry out assassinations. An automated artillery system could fire into a city and decide when targets are available. The scale of potential harm is enormous.

Ahead of the meeting, the International Committee of the Red Cross, the organization that serves as guardian of the Geneva Conventions, issued an urgent appeal. It said legally binding rules should be drafted within a year. It argued that the moral and legal risks of autonomous weapons are too great to allow an extended period of unrestricted development. The ICRC’s call was backed by the U.N. secretary-general and by Pope Leo. Yet that urgency was not reflected in the official outcome.

Instead of a one-year negotiating mandate, the compromise reached in Geneva gives countries two years to propose regulations. Even that modest step was qualified: the final agreement reportedly leaves room for some of the future regulations to remain “optional.” In other words, states may be allowed to pick and choose which constraints they accept. For many observers, an optional regulation is a contradiction in terms. A rule that does not apply to everyone is not a rule at all; it is a suggestion.

The two-year schedule is also likely to be optimistic. Even after proposals are made, negotiations on which regulations to adopt and how to enforce them will probably continue for years. Every round of procedural discussion brings the technology further onto the battlefield. Humanitarian organizations say the delay is a deadly gamble.

Semi-autonomous weapons and drones have already been responsible for a significant number of civilian deaths in conflicts around the world. In Ukraine, drones are now a constant feature of combat, used for reconnaissance, artillery correction, and direct strikes. Some of these systems have enough autonomy to track a target or maintain a flight path even if communications with an operator are cut. The result is that mistakes can happen at machine speed, and those mistakes can kill civilians.

In Gaza, concerns have been raised about the use of artificial intelligence to help select targets. The Israeli military has acknowledged using AI-based systems to assist with targeting, and the widespread destruction and civilian casualties in Gaza have led to calls for closer scrutiny of how automated tools are used in the fighting. The idea that an algorithm could decide, even indirectly, which buildings or people are legitimate targets is deeply troubling to international lawyers and human rights groups.

In Sudan, too, drones and automated systems have been linked to violations of international humanitarian law. The civil war between the Sudanese Armed Forces and the Rapid Support Forces has produced a horrific toll on civilians, and the use of aerial bombardment and drone strikes in populated areas has made the situation even worse.

These examples underline the urgency of the issue. When a drone or an autonomous system is used in an attack, it is not always clear who is responsible. Was the target chosen by a human, or by a computer? Did the operator foresee the possibility of civilian harm? If an autonomous weapon violates the principles of distinction, proportionality, and precaution, who should be held accountable? These are fundamental questions under international law, and none of them can be answered by vague, optional guidelines.

The Geneva talks were supposed to address those questions. Instead, they were largely derailed by procedural objections and geopolitical resistance. According to NPR, Russia, Israel, and the United States were the primary obstructions. On the first day, Russia delayed the proceedings by objecting to the meeting’s agenda. It was a familiar tactic. Russia has historically resisted international constraints on military technology and has opposed attempts to create binding rules that could limit its own arsenal.

The United States, for its part, proposed a five-year window for discussing nonbinding, optional guidelines for autonomous weapons. Washington argues that the technology is still developing, that a rush to regulate could stifle innovation, and that existing laws of armed conflict are sufficient. But critics point out that the same arguments have been made for decades, with little progress. The American position would effectively push meaningful regulation far into the future, perhaps indefinitely.

Israel, too, has been viewed as an obstruction. Israel has deployed advanced targeting systems in recent conflicts, and it has been reluctant to accept new legal constraints that might restrict its military freedom of action. According to NPR, Israel was among the countries that slowed progress toward a binding agreement.

The outcome is especially frustrating given that the broader diplomatic framework already exists. The Convention on Certain Conventional Weapons has been in force since 1983, serving as the platform for limits on land mines, booby traps, incendiary weapons, and blinding lasers. Since 2013, states have been discussing lethal autonomous weapons systems under this convention. In 2019, they agreed on eleven guiding principles, but those principles are not legally binding. The central issue remains unresolved: should fully autonomous weapons be banned outright, or simply regulated?

Many countries and civil society organizations insist that there is a grey zone of semi-autonomous systems and that any new treaty must cover the full spectrum of autonomy in weapons. They point to the work of campaigns such as the Stop Killer Robots coalition and the International Committee for Robot Arms Control, which for years have called for a strong, legally binding instrument. Pope Leo’s support, along with that of the U.N. secretary-general, gave the issue moral weight. But moral weight, it seems, is not enough to overcome the strategic interests of major military powers.

The future of autonomous weapons is not a distant science fiction scenario. It is happening now. There are missiles that can select and engage targets on their own. There are loitering munitions that can circle over an area and strike based on an algorithm. There are battlefield management systems that process intelligence and recommend targets faster than any human. The technology will continue to develop, with or without international rules. The question is whether those rules will keep pace.

For now, the only clear achievement is a promise to keep talking. For diplomats, that may look like progress. For the ICRC and the millions of civilians living in conflict zones, it looks like another missed opportunity to protect human life. The people most affected by autonomous weapons are unlikely to feel reassured by a two-year proposal period or by the possibility that any eventual rules will be optional.

The next round of negotiations will come and go. There will be more proposals, more objections, more compromises. Eventually, some document may be agreed. But if it remains nonbinding, voluntary, and optional, it will do little to stop the spread of killer robots. And the wars in Ukraine, Gaza, Sudan, and elsewhere will continue to show the world what happens when machines are allowed to make life-and-death decisions without adequate oversight.

Share This Article
mediawatchbot
11 Min Read