Palantir Technologies Inc. (PLTR) CTO Shyam Sankar is pushing back on the idea that his company's AI software had anything to do with a strike that killed at least 123 schoolchildren on Feb. 28 in Minab, Iran.
Sankar made his case during an interview on CBS "Face the Nation" on Sunday, where the conversation turned to AI-assisted military targeting. His pitch: AI can speed up military planning, even for operations involving 1,000 strikes, while helping personnel review no-strike lists and flag locations near civilian infrastructure.
Asked about a report alleging that personnel leaned too heavily on Palantir's Maven system during attacks in Iran, Sankar called the account "inaccurate" and urged the public to wait for the final report.
The CTO said "the likely cause is human error," though he stopped short of offering a definitive conclusion. Sankar also said Palantir quickly developed additional AI agents to independently check information and reduce the risk of another incident.
Human Error, AI Risks Fuel Military Debate
The strike has drawn scrutiny well beyond Palantir's software. A March report said an ongoing U.S. military investigation pointed to a targeting mistake involving a Tomahawk missile that struck Minab's Shajarah Tayyebeh elementary school on Feb. 28. That account suggested outdated data and basic human error, rather than AI, may have caused the deadly strike.
The debate arrives as military organizations expand their use of autonomous and AI-enabled systems. Last week, Sarah Yager, OpenAI's first human rights and responsible deployment lead, warned that "everyone is very worried" about AI's role in armed conflict and autonomous weapons.
Yager said human rights and international laws governing warfare should serve as the baseline for AI use in military scenarios. She warned against letting AI companies define ethical standards themselves, arguing that industry-led oversight could lead to unchecked practices.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by MarketDash editors.
