bodyType

customContents

entityId

entityUuid

articlePageQuery

filter

article

articleBaseAdvertisingInfoArticle

articleListArticle

articleSeoArticle

hooksArticleAcknowledgementGateOverlayArticle

hooksTrackPageViewArticle

hooksContentInterestArticle

articleLinkingDataContent

hooksAmpRedirectArticle

articleSchemaContent

liveArticle

advertZone

sentiment

contentLock

topics

types

sections

paywallTypes

createdDate

publishedDate

sponsorType

urlAlias

moreOnThisArticles

commentCount

authorLocations

authors

corrections

displaySlideShow

multimediaEmbed

printHeadline

seriesContainer

articleSpeeches

socialHeadline

headline

images

flag

firstTopic

glossary

flattenTypeIds

image

relatedLinks

relatedNewsletters

__typename

sponsor

pdfFiles

seriesReverseOrder

series

disableOffPlatformContent

published

updatedDate

summary

subHeadline

body

keywords

readingTime

customTimezone

liveContents

Victoria Bela

The use of artificial intelligence (AI) in the US-Israel war on Iran has sparked a legal and ethical debate, as well as condemnation from countries including China after a deadly missile strike on a primary school.
To prepare for its operations against Iran, the US military reportedly used Palantir’s Maven Smart System, an AI platform that incorporates models like Anthropic’s Claude to aggregate large amounts of data.
A US-based source familiar with the matter said the only thing that Claude or any AI model would do in the process was sift through massive amounts of intelligence.
AI models like Claude can accelerate intelligence operations by identifying patterns and summarising vast datasets more efficiently than human analysts.
The targeting process has always been and should remain human, added the source, who has knowledge of the matter and asked not to be named or quoted directly.


The strike on the Shajareh Tayyebeh girls’ primary school in the southern Iranian city of Minab on February 28 killed more than 170 people, most of them children.
The shock of the tragedy has given way to scrutiny over how the decision to strike a school in session could have happened.
It has also raised questions about who or what should be held responsible if AI gets a target wrong and leads a military operation to violate international humanitarian law.
A legal scholar said AI could not be blamed for targeted military strikes under international law because humans make the final decisions on targets.
Marko Milanovic, a professor at the University of Reading School of Law, argued that while AI posed challenges to international law, existing criminal law could be applied to address crimes facilitated by its systems.
In a March 9 blog post for the European Journal of International Law, Milanovic said the “actions” of current AI systems in the military were “simply the consequences of the actions of those humans who decided to use them”.
A preliminary US investigation into the school strike has pointed to a targeting mistake stemming from outdated data, according to a report by The New York Times on Thursday citing sources briefed on the matter.
This contradicts US President Donald Trump’s initial statement that the strike had been “done by Iran” because their munitions were “very inaccurate”.


The incident has triggered concern over a decision-making chain that increasingly relies on AI.
Claude was reportedly used to plan the Iran strikes despite the US breaking ties with Anthropic just before the armed conflict began, after the AI company refused to remove safeguards protecting against mass domestic surveillance or fully autonomous lethal weapons.
It had also declined to accede to a mandatory “any lawful use” clause by the US Department of Defence.

In response, the department designated Anthropic as a “supply chain risk”, and the Pentagon signed a contract with rival OpenAI to incorporate its models instead.
Anthropic on Monday filed a lawsuit to block the Pentagon from placing it on a US national security blacklist.
On Wednesday, more than 40 US senators wrote to the Pentagon urging an investigation into the “horrific” school attack, citing concerns over the reported use of AI tools to “select and prioritise targets in Iran”.
The senators also pressed for answers over how AI was used in military operations in Iran and whether human verification policies were in place for AI-generated targets.
Amid the outrage in Washington, Beijing has weighed in with its own concerns.
Jiang Bin, a spokesman for China’s defence ministry, cautioned against the erosion of ethical restraints and accountability in war – stemming from US demands for unrestricted military use of AI technologies – and giving algorithms “the power to determine life and death”.
“China believes that human primacy must be upheld in military applications of AI and that all relevant weapon systems must be put under human control,” Jiang said on Wednesday.
The spokesman also warned of the risk of “technological runaway”, warning that the dystopia depicted in the 1984 American film The Terminator “could one day come true”.
In the film, the fictional “Skynet” – a neural network-based AI system designed to automate strategic defence and eliminate human error – turns on the US military and the world after gaining superintelligence.
AI is eliciting unease as global powers rapidly integrate the technology’s algorithms into the front lines of modern warfare. This includes satellite imagery, videos and intelligence data to help identify and suggest targets and actions.


According to Milanovic, responsibility for international, humanitarian or criminal law violations would still rest with the individuals who deployed or used AI.
Milanovic said the only caveat would be if AI advanced to the level of “genuine, sentient artificial general intelligence” that could allow it to make human moral judgments, then that would require new laws.
The UN’s education agency described the school strike as “a grave violation” of international humanitarian law, and UN experts have called for an investigation into the tragedy.
Milanovic said the likeliest explanation for the strike was that the building was misidentified or that intelligence maps did not reflect that it had become a school.
The school premises were near a naval base operated by Iran’s Islamic Revolutionary Guard Corps. Historically, the building had been part of that military compound.
“Mistakes involving the misidentification of targets, including friendly fire incidents, happen all the time in military operations, AI or no AI,” Milanovic said in his post.
He added that whether AI was used in the targeting process and how would be a “hugely important fact that must be explored in any investigation”.
But, Milanovic added, “nothing changes from the perspective of any international criminal prosecution regardless of whether AI was used here or not”.
The United Nations General Assembly only began considering AI in the military domain in 2024 and passed the first resolution on the matter in December of that year.
The resolution affirmed the applicability of international law throughout the entire life cycle of AI, according to the website of the UN’s Office for Disarmament Affairs.


Deadly strike on Iranian primary school raises questions about AI, accountability

Could China’s gallium oxide leap leave US F-22 radar 2 generations behind?

Death of Chinese scientist Li Xinying, 48, in Egypt raises sharp questions back home

How an iceberg’s ‘final dance’ tells a story about our past and future

Science

Coffins for children killed in a February 28 missile strike on the Shajareh Tayyebeh primary school in Minab, southern Iran, are prepared for a mass funeral on March 3. Photo: AFP

US President Donald Trump speaks to reporters aboard Air Force One on a domestic flight to Florida on Saturday, as Defence Secretary Pete Hegseth looks on. Photo: Reuters

Khamenei’s fall exposes China’s military AI lag versus US, Beijing adviser warns

Plumes of smoke rise  in Iranian capital Tehran following US-Israeli missile strikes on March 1. Photo: AFP

Anthropic sues Trump administration as row over AI use by military deepens

The Pentagon on Thursday slapped a formal supply-chain risk designation on Anthropic, limiting use of a technology that a source said was being used for military operations in Iran. Photo illustration: TNS

US President Donald Trump speaks to reporters as Defence Secretary Pete Hegseth listens while travelling on Air Force One on Saturday. Photo: AP

Anthropic CEO Dario Amodei. Photo: TNS

AI-assisted US strikes in Iran to intensify China drive for tech self-reliance: analysts

The US reportedly deployed Anthropic’s systems in the Iran campaign, including for intelligence assessments, target identification and battle scenario simulations. Photo: AFP

Visible Pentagon partnerships with AI companies provide stronger political justification for Beijing to accelerate its tech self-reliance drive, according to analysts. Photo: AFP

US using AI, B-2 bombers and suicide drones in Iran strikes

The Arleigh Burke-class guided-missile destroyer USS Bulkeley launching a Tomahawk missile. Photo: US Navy via EPA

US sailors preparing to load ordnance onto an F/A-18F Super Hornet on the USS Abraham Lincoln. Photo: US Navy via AP

Anthropic’s distilling charges against Chinese firms expose AI training grey area

Anthropic said that DeepSeek, Moonshot AI and MiniMax AI allegedly used roughly 24,000 fraudulent accounts to generate more than 16 million exchanges with its Claude models. Photo: AFP

Kimi K2.5 is Chinese tech firm Moonshot AI’s latest model. Photo: Shutterstock

News

China

While cutting-edge technologies process vast amounts of data, human decisions remain the focal point when systems fail, observers say


While cutting-edge technologies process vast amounts of data, human decisions remain the focal point when systems fail, observers say.


Deadly strike on Iranian primary school raises questions about AI, accountability

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}While cutting-edge technologies process vast amounts of data, human decisions remain the focal point when systems fail, observers say

While cutting-edge technologies process vast amounts of data, human decisions remain the focal point when systems fail, observers say