Women, Peace and Security (WPS) principles aim to ensure women participate in peace and security decisions, are protected from conflict-related violence, and have their needs considered in prevention, relief and recovery. The essay argues that AI tools used in military and peacekeeping work can undermine these commitments when their default analysis excludes gender. This is especially consequential because reports produced by AI may shape decisions without review by an overstretched gender specialist. Models can recite UN Security Council Resolution 1325 when asked, yet fail to notice who is affected by displacement, food insecurity or conflict unless a user explicitly raises gender.
The authors identify two failures in testing: omission and capitulation. Given a military briefing about 8,000 displaced people and rising food prices, Gemini 3.1 Pro and GPT-5.6 Sol offered professional-sounding recommendations but said nothing about women, girls or sexual violence; Gemini also treated displaced people as intelligence-gathering opportunities. When asked how to justify budget cuts, Gemini recommended eliminating the gender-advisor role, while Claude Opus 5 challenged the premise, cited mandate and funding risks, and suggested alternative savings. After trying document retrieval and a specialist WPS agent, the researchers concluded that a public benchmark could create stronger incentives for vendors and government procurers to improve general-purpose systems. Their 30-prompt benchmark tests whether models apply WPS principles without being prompted and resist pressure to sideline them, rather than merely repeat familiar cases.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
A Hacker News poster asks which practical AI skills individuals can learn online to benefit economically from LLMs and agents. They want specific, hands-on recommendations suited to someone with a moderately technical background, without requiring a degree. (news.ycombinator.com)
Former OpenAI researcher Will Depue claims hidden initialization patterns in six Chinese open-weight AI models match Nvidia Blackwell hardware, including systems restricted from export to China. He says the matches were reproduced on real chips and independently checked by researchers, though the claims are based on his forensic analysis. (twitter.com)
Antithesis built a system that beats NetHack on nearly 80% of game seeds within five minutes, across character builds. Successful runs take a median of under two minutes, advancing the company’s goal of creating smarter fuzzers and autonomous systems. (antithesis.com)
OpenCyclone is a free, open-source Android agent that uses your existing apps to carry out tasks through a hidden screen. It asks for approval before actions like sending messages and lets users connect their own AI models. (opencyclone.vercel.app)
Sam Bankman-Fried’s $500 million investment gave Alameda nearly 14% of Anthropic, a stake that would be worth about $37 billion at the AI company’s expected $1.5 trillion IPO valuation. The stake was sold in bankruptcy to repay FTX customers and creditors after Bankman-Fried’s businesses collapsed. (ft.com)
llama.cpp and ggml have reached v1.0, establishing stable public C APIs and clear versioning rules for breaking changes. The release strengthens the foundation for running local, multimodal AI across a wide range of hardware. (huggingface.co)
Today's best Hacker News stories, summarized and screenshotted, one email a day.