Autonomous-weapons talks set 2028 target, leave treaty open | Magica
Autonomous-weapons talks set 2028 target but leave binding rules open
Editorial Team
••📖6 min read
UN-backed talks agreed on two more years of work toward an autonomous-weapons instrument, without committing to a treaty. Consensus kept major military powers involved while allowing opponents to weaken the mandate.
Countries agreed to work toward an autonomous-weapons instrument by 2028, leaving its legal force undecided.
Consensus rules gave opponents leverage despite majority support for a treaty.
A binding protocol remains possible; no new prohibitions or safeguards were adopted.
UN-backed talks in Geneva ended on November 21 with an agreement to develop an international instrument on autonomous weapons over the next two years, but without a commitment that it would impose binding obligations. The compromise preserved participation by major military powers while falling short of the treaty mandate sought by many states and campaigners.
The stakes are control over machines that can select and attack targets without human intervention after activation. Campaigners say such systems are already appearing in conflicts. The agreement establishes further work, rather than rules governing their use.
Guterres set a 2026 deadline for launching negotiations. The Geneva agreement targets a possible accord by 2028 without committing states to binding rules. Source: Associated Press.
Consensus preserves participation and limits the mandate
The week-long review conference took place under the Convention on Certain Conventional Weapons, or CCW, a treaty framework for prohibiting or restricting weapons deemed excessively injurious or indiscriminate. Its 128 parties make decisions by consensus, so any single country can block agreement. The mandate ultimately asked states “to elaborate by consensus” an instrument that might carry no legal obligations, according to the account of the negotiations.
Austria, Belgium, Brazil, Ireland, Mexico, New Zealand and Panama had sought negotiations on a binding treaty within a year. Stop Killer Robots, a global coalition of more than 300 organisations campaigning for restrictions, said over 95 states supported a treaty, but a small number weakened the mandate in the final hours, according to the conference report. That is the coalition's count of support, rather than a recorded vote on the final agreement.
Employee equity at Anthropic’s IPO price is reportedly worth $232 billion, while book rightsholders await a first settlement installment estimated at $2,203.56 per work. Vesting, shared copyrights and founder voting rights complicate the comparison.
Editorial Team
Diplomats involved in the talks accused Russia of deliberate stalling. Their account distinguished that conduct from US delegates' efforts to secure a consensus statement, even as Washington resisted committing to a binding treaty. The same report identifies the UK, France, Germany, Japan, South Korea and Turkey as defence technology powers exploring AI-enabled weapons that also advocated a nonbinding agreement.
Some participants argued that existing international humanitarian law was sufficient. Ukraine, China and Iran were among countries expressing support or willingness for a new protocol. The disagreement therefore extended beyond Russia and the United States, and support for a protocol did not follow a simple division between geopolitical blocs.
Participation itself was contested: the exclusion of several nonmember countries prompted a protest in the chamber on the final day, the account says. A consensus among the convention's parties does not mean all states seeking a say were included in the negotiations.
The CCW has produced binding additions before. The UN's treaty record shows that its Protocol V on explosive remnants of war was adopted on November 28, 2003, and entered into force on November 12, 2006. That precedent demonstrates the framework's capacity to add law, without establishing that states will agree on autonomous weapons.
There is also precedent for working outside the CCW: the negotiations report points to the 2008 cluster-munitions ban as a treaty states pursued through another route. This week's agreement keeps the autonomous-weapons process within the consensus framework.
The UN’s Protocol on Explosive Remnants of War illustrates the CCW’s record of binding additions. Protocol V was adopted in 2003. Source: United Nations Treaty Collection.
The question is who controls the use of force
The International Committee of the Red Cross, or ICRC, is a neutral, independent humanitarian organisation protecting people affected by armed conflict and promoting humanitarian law. It was at the origin of the Geneva Conventions and has urged internationally agreed limits on autonomous weapons since 2015.
Its May 2021 policy position explains the mechanism: after human activation, a system initiates force in response to sensor information and a general target profile. The user does not choose, or even know, the specific target and precise timing or location of the strike. The organisation identifies risks to civilians and combatants, challenges for compliance with humanitarian law, and ethical concerns about delegating life-and-death decisions to machines.
This distinction matters when assessing deployment claims: using AI to help identify targets does not by itself establish that a weapon selects and strikes them without human intervention.
Stop Killer Robots says autonomous systems have been tested or deployed, including in Ukraine. Separately, reporting on the outcome records activists' assessment that drones capable of independent strikes have appeared in selected instances in conflict zones, with widespread rollout yet to happen. These are attributed assessments, not a comprehensive inventory of weapons in service.
The ICRC's proposals show what substantive regulation could involve. It recommends legally binding prohibitions on unpredictable autonomous weapons and systems designed or used to apply force against people. For other systems, it proposes limits on targets, duration, geography and scale, restrictions on use where civilians are present, and requirements for effective human supervision, timely intervention and deactivation. Those are the organisation's recommendations, not safeguards adopted in Geneva.
The next decision is whether the instrument will bind states
Swiss diplomat Julien Thoni, who presided over the review conference, said securing the major military powers' participation was a significant step. He described a process in which countries would work on an instrument and then decide whether it should be a treaty and how far its provisions should bind them, according to his account of the agreement.
The timetable remains disputed. Critics quoted in the negotiations report pointed to 13 years of discussion without the rules they sought. UN Secretary-General António Guterres had set a 2026 deadline for launching negotiations toward an agreement. Yet producing the instrument within the next two years will itself be demanding, according to Izumi Nakamitsu, the UN high representative for disarmament affairs.
Nakamitsu said governmental experts would recommend the instrument's legal nature and specific form, including whether it should be a binding protocol. She stressed that a legally binding outcome remained possible, while warning that the timetable would require substantial negotiations.
The next decisions concern both which weapons to prohibit or regulate and whether those provisions will create legal obligations. The 2028 target is a timetable for that work, not a guarantee of a treaty or of enforceable limits on autonomous force.
Moore Threads is preparing a December 29 reveal for Lushan, a desktop GPU pitched at gaming and local AI inference. Its claimed memory record remains unquantified, and capacity alone will not establish performance or value.
Amazon is in advanced talks to buy Decart, whose software helps AI models run efficiently across chips. The reported deal could strengthen AWS’s Trainium business, with conditional payments and five-year founder commitments making execution central to its value.
The University of California is expanding ChatGPT and Codex access across 10 campuses, with extra charges for advanced models and an unresolved dispute over faculty authority.
Ray-Ban’s US store cuts the Gen 2 Wayfarer by 25% to $284.25. Buyers still face separate prescription-lens prices, a cheaper Gen 1 option and newer glasses outside Meta’s frame sale.
Ubisoft has made the team behind its Teammates experiment a dedicated subsidiary targeting AI characters with memory and outside game creators. Earlier playtests show both progress and limitations, while commercial plans remain undisclosed amid group cost cuts and refinancing.
Neuralink has shown a trial participant controlling GTA VI through intended movements. Earlier gaming use and company-reported trial results put the demonstration in context, while its reliability across users remains unmeasured in the reports.
SoftBank founder Masayoshi Son reportedly explored financing a $1 trillion acquisition of memory-chip maker SK hynix, but the company says the deal is no longer under active consideration. The target brings surging earnings, existing shareholders and customer commitments that complicate any simple equation between ownership and access to AI-memory supply.
Nvidia reported $115.8 billion in quarterly revenue, but its shares fell 14% Friday amid a wider technology selloff, a leadership transition and a reported accounting investigation. Strong cash generation offers a counterweight to concerns about the durability of AI spending.
The Nasdaq 100 fell 5.6% on November 20 and closed 19.8% below its record after an intraday bear-market breach. Nvidia’s revenue surge and Anthropic’s huge IPO did not halt the rout; infrastructure bills and resilience elsewhere in the market complicate the picture.
Tesla is sending FSD v15 to its first customers, adding a visualization of planned driving and changes to navigation and hazard detection. The software remains supervised, and its claimed 25-fold intervention improvement lacks a disclosed baseline.
OpenAI reports sharply lower reasoning-extraction rates in internal GPT-6.2 tests and backs sanctions against illicit foreign distillation. Its own research and rival Anthropic’s experience show why protecting hidden reasoning leaves capability transfer and customer costs unresolved.
Alberta proposes 600-metre setbacks for large data centres and expanded utility oversight, but exemptions and rules still to be written will determine how much protection communities receive.
President Donald Trump backed Commerce Secretary Howard Lutnick as lawmakers scrutinize alleged delays in restricting a suspected route for AI chips to China. A new congressional inquiry also questions Nvidia’s influence on an existing investigation.
Amazon’s month-old Echo Time is $50 off for Black Friday. Its touchscreen, home hub and optional sleep-aware alarm cost $90 more than the reported Echo Spot price, while Alexa+ access and evidence of better waking remain key buying questions.
Merck will pair its experimental data with Isomorphic Labs’ AI to seek small-molecule cancer drug candidates. The deal follows Isomorphic’s $5 billion financing, but its payments, target and development timetable remain undisclosed.
Thinking Machines Lab released a roughly 20GB open-weights model that reuses its middle layers for reasoning. It brings the Inkling family to local hardware, with competitive coding results but workload-dependent costs and preliminary evidence on oversight.
Apple’s first HomePod mini update since 2020 adds Mac audio and four-speaker surround, but Siri AI remains unavailable, core home-hub features carry over, and reviews disagree on playback reliability.
Treasury Secretary Scott Bessent says an unnamed Anthropic founder will announce a large gift to children's investment accounts within weeks, but the amount, recipients and connection to existing philanthropic pledges remain undisclosed.
Google says air cooling will limit water use at its Visakhapatnam AI hub, but environmental proceedings and residents’ accounts raise broader questions about land, electricity, public incentives and the evidence behind promised benefits.