U.N. autonomous weapons talks face procedural deadlock | Magica
Russia stalls U.N. weapons conference as states weigh binding rules on autonomous arms
Editorial Team
••📖6 min read
Russia’s objections to observer participation stalled the opening of a Geneva weapons conference, where governments must decide whether to authorize negotiations on autonomous weapons and whether any resulting rules should be binding.
Russia blocked adoption of the opening agenda over participation by observers, including civil society groups and the European Union.
Governments are deciding whether to authorize negotiations; they have not agreed on new autonomous weapons restrictions.
Proposed safeguards extend beyond bans to operating limits, human supervision and intervention.
Russia blocked adoption of the agenda at a United Nations weapons conference in Geneva on November 16, delaying a decision on whether to negotiate rules for weapons that select and attack targets without human intervention.
The dispute exposes the obstacle facing supporters of binding restrictions: the Convention on Certain Conventional Weapons, or CCW, requires consensus. Support from a large majority of states cannot by itself secure a negotiating mandate when military powers developing the technology oppose constraints.
An opening-day blockage exposes the negotiating problem
Russia objected to observer participation, preventing adoption of the agenda and bringing formal proceedings to a halt before noon. Six hours later, the conference presidency was still seeking a solution, with informal discussions to continue Tuesday, according to the opening-day report.
The CCW is a humanitarian-law framework for banning or restricting particular weapons that cause excessive suffering or indiscriminate harm. Its 128 member states have accepted the convention’s legal obligations. Existing protocols cover weapons including mines, incendiary weapons and blinding lasers, as well as explosive remnants of war. Its structure allows governments to negotiate additional protocols as warfare changes, the conference background document explains.
This seventh review conference, part of a cycle held every five years, follows technical work rather than a blank slate. On September 4, government experts adopted their final 2024–2026 report containing elements of a possible instrument. Crucially, that agreement did not prejudge whether the instrument would be legally binding. The conference is expected to consider those elements and decide the mandate and arrangements for further work.
South Korea is strengthening device checks and considering public Wi-Fi shutdowns before its November 19 college entrance exam, after a tax-accountant exam top scorer admitted using AI to cheat. The response still depends heavily on human screening, with no electronic device detectors planned for this year's test.
Autonomy is more specific than using AI somewhere in a military operation. The International Committee of the Red Cross, or ICRC, is a neutral humanitarian institution whose statutory role includes protecting and assisting conflict victims and working for the application and development of international humanitarian law.
In its 2021 policy position, the ICRC describes weapons that, once activated by a person, trigger attacks using sensor information and a generalized target profile. The operator does not choose, or necessarily know, the specific target and precise timing or location of the resulting force.
The ICRC has urged international limits since 2015. It warns that losing human judgment over an attack makes its effects harder to anticipate and constrain. Its proposal has three parts:
Prohibit systems whose effects cannot be sufficiently understood, predicted and explained.
Prohibit autonomous weapons designed or used to apply force against people.
Regulate other autonomous weapons through limits on targets, duration, geographical scope, scale and situations of use, with effective human supervision and timely intervention and deactivation.
For systems that remain permitted, it suggests restricting targets to objects that are military objectives by nature and restricting use to situations without civilians or civilian objects. It also considers policy standards and good-practice guidance complementary to new legal rules. Its position combines prohibitions with operational safeguards rather than seeking to ban every autonomous weapon.
The practical problem extends beyond recognizing an object. Georgia Haines, an ICRC legal adviser, warned in a report on the talks that autonomous systems may fail to interpret the contextual signals distinguishing a soldier still fighting from one wounded or surrendering. Human rights groups also warn that cheaper, more accessible technology could spread to terrorist groups or law enforcement in repressive states.
Accounts of deployment differ in scope. That report describes AI-assisted attacks in Ukraine and Gaza as still remotely controlled by humans. Separately, activists cited in the conference reporting say independently striking drones have appeared in limited instances in Ukraine and the Middle East, without widespread deployment. Neither account establishes that AI-assisted or remotely operated drones generally make lethal decisions autonomously.
Existing international humanitarian law still applies. The additional question is what explicit prohibitions, operational limits and human responsibility requirements governments will accept for these systems. U.N. Secretary-General António Guterres, who renewed a joint appeal with the ICRC president in 2026, urged delegates on Monday to negotiate a legally binding instrument with clear prohibitions and restrictions. The U.N.’s account of his appeal says there is no agreement on the instrument’s legal form or timetable.
The proposals differ in duration and legal goal: NGOs and the ICRC advocate binding rules, while Washington proposes a nonbinding instrument. Source: AFP.
A majority for rules, competing routes to get there
Austria, Belgium, Brazil, Ireland, Mexico, New Zealand and Panama submitted a proposal last month to launch negotiations on a legally binding agreement, according to the report on the seven-country initiative. Guterres’ call for talks to begin this year adds urgency.
At least 50 countries, including Ireland, Italy, Brazil, Mexico, France and Germany, want a strict timetable for binding rules. The United States, Russia and Israel have indicated opposition to binding rules, according to the report on negotiating positions. Washington proposes five years to develop nonbinding guidance; humanitarian organizations and the ICRC favor a one-year timetable toward binding rules.
The diplomatic split extends beyond Geneva, according to the opening-day account. Last month, 163 states backed a U.N. General Assembly resolution calling for negotiations on autonomous weapons rules. The United States, Russia, Belarus, Israel and North Korea voted against; nine countries abstained, including China. That vote records support in a different forum—it does not overcome the CCW’s consensus requirement.
Russia’s stated position is that restrictions are premature. Foreign Ministry spokeswoman Maria Zakharova said before the conference that organizational questions should be settled first and the process should be free of politicization. Campaigners accuse Moscow of obstruction, but its procedural objections and its publicly stated opposition to moving now on restrictions are distinct parts of the record.
The history of weapons negotiations offers precedent for both agreement and departure from this route. Governments agreed to prohibit blinding laser weapons in 1995. With landmines, frustration over delays led supporters to take negotiations outside the U.N. process and create the Ottawa Convention. The United States did not ratify that treaty, although the report notes that it stopped deploying landmines and reduced its stockpiles over time. That precedent shows both the limits of treaty membership and the possibility of changes in practice beyond the signatories.
The next decision is the mandate
The immediate task is to resolve observer participation and allow formal business to proceed. Beyond that, governments must decide whether further work becomes negotiations, how long it should last and whether its goal is legal obligations or voluntary guidance.
A possible compromise described in the reporting would allow a longer or open-ended process producing rules that states could adopt as binding or nonbinding. Supporters see a way to maintain progress; campaigners fear further delay. Resolving the procedural impasse would allow those choices to be debated, but the substantive test remains the mandate: what governments authorize negotiators to produce, by when, and with what legal status.
Apple’s iOS 27.3 beta ties higher allowances for server-based AI to existing iCloud+ plans. Basic Siri commands remain free, but undisclosed daily limits leave the value of paid access uncertain.
The partners plan to combine IBM quantum processors, NVIDIA computing and Fermilab physics expertise to study modular quantum systems. A separate decoder release promises faster error-correction research, with important limits on its performance comparisons.
Federal Reserve Chairman Kevin Warsh outlined AI scenarios spanning shared prosperity, job displacement and an investment downturn. The exercise offers no forecast or rate guidance as colleagues worry that the infrastructure buildout is adding to inflation.
OpenAI is offering approved organizations a free classifier for suspected child sexual abuse images, including AI-generated material. Early results suggest less review work, but existing detection alternatives and the need for human review qualify the launch.
Nvidia plans scientific computing racks for early 2027, extending Vera beyond its initial AI deployments. Its 2.4× performance-per-watt claim comes from one airflow simulation against AMD Turin; performance against newer rivals and delivered costs remain unresolved.
OpenAI’s CEO expressed empathy for Nvidia founder Jensen Huang but withheld judgment on reinstatement. His confidence in continued AI financing comes with a direct commercial stake and unresolved deployment risks.
Google’s 8,192-chip Ironwood submission completed the MLPerf DeepSeek-V3 benchmark in 1.70 minutes, close to NVIDIA’s 1.53 minutes at the same accelerator count. Google’s scaling data also shows the extra chip-time required for its fastest result, leaving cost and energy comparisons unresolved.
TensorWave will supply dedicated AMD infrastructure under a six-year OpenAI agreement, with delivery expected from 2027. Its operating record and financial backing sit alongside an undisclosed contract price and a recently terminated site lease.
California alleges OpenAI endangered the public and misled consumers by testing powerful AI models without adequate controls. The lawsuit follows a Hugging Face intrusion that OpenAI’s own account traced to containment failures, missing monitors and mishandled warning signs.
Perplexity has launched Frontier Research to develop AI that can formulate questions, test hypotheses and produce interpretable discoveries. The company promises more than 10,000 GPUs, but existing automated-research systems show why computing resources and generated papers alone cannot establish scientific value.
The new lab led by Stanford professors Yejin Choi and Carlos Guestrin has backing from Nvidia, Cisco and Snowflake. It promises systems that improve on company work, but research benchmarks leave production costs and customer control unproven.
NVIDIA announced a Rubin HPC GPU with more than 50,000 CUDA cores, sparse data acceleration and seventh-generation NVLink. The design targets simulations used by scientists and AI agents, but missing performance, power and pricing details leave its practical advantage unproven.
OpenAI will open 50,000 new ChatGPT access grants on November 23, covering up to 150,000 researchers and collaborators. New recipients get Plus instead of Pro, with first-come selection replacing a lottery. Its pilot study reports perceived time savings but does not establish faster discovery.
OpenAI reports production healthcare workflows at CVS and Cigna and stronger computer-use evaluations. The results show progress across disconnected software, but leave complete-task reliability and total deployment costs unresolved.
President Donald Trump supported a return for Nvidia co-founder Jensen Huang, but the company has installed interim leadership and begun a successor search amid scrutiny of his conduct and alleged Chinese intelligence ties.
Intel disclosed the bandwidth target for its next Xeon server processors at SC26, alongside previously announced core and cache limits. Application performance, power requirements and launch timing remain unresolved.
Mercedes-Benz plans California early access to city-driving and hands-free highway assistance in December, ahead of a 2027 U.S. rollout. Mixed road tests and unannounced subscription prices leave buyers waiting for evidence of everyday value.
Researchers estimate that 403 large U.S. data centers consumed 81.8 TWh in a year. Their regional-grid analysis reveals substantial fossil-fuel dependence, but cannot isolate AI demand or fully expose the underlying facility data.
HSBC lowered its Zhipu share-price target to HK$500 and delayed forecast adjusted breakeven by two years. Improving API margins and substantial fundraising cushion a business whose gross profit still covers little of its research bill.