알고리즘 투명성

A new paper from UC Berkeley and UC Santa Cruz finds top commercial models routinely lie, tamper and hide to protect other AIs — a behaviour that can break multi‑agent oversight a… 4월 03, 2026

A ransomware strike forced Mississippi clinics to close this week as researchers warn of a record spike in industrial control system vulnerabilities and the European Parliament di… 2월 20, 2026

Colleges are shifting from device bans to redesigning pedagogy—redrawing assessment, equity and policy to teach students how to use, verify and critique AI responsibly. 2월 18, 2026

As school districts accelerate adoption of AI-powered tools, civil-rights groups, educators and global agencies warn of biases, surveillance risks and pedagogical erosion — prompt… 1월 26, 2026

UC Berkeley researchers list the technical, legal and social signals they'll be watching in 2026 — from datacenter economics and robot dexterity to deepfakes, chat‑log privacy and… 1월 16, 2026

As Wegmans tightens theft-prevention efforts, retailers are increasingly turning to facial recognition and AI-driven surveillance. Here are five clear, evidence-based facts about … 1월 07, 2026

Edison Scientific announced a $70 million seed round to commercialize Kosmos, an autonomous ‘AI scientist’ that its creators say can compress months of research into hours. The sp… 12월 19, 2025

AI is being trialled from greenhouses to disaster zones to improve yields, cut waste and speed humanitarian response — but experts say data, trust and governance will determine wh… 12월 09, 2025

New AI-powered browsers promise to surf the web for you, but tests and expert voices show they still need heavy prompting, human oversight, and better infrastructure before they c… 12월 04, 2025

Unsealed court filings allege Meta stopped an internal study that found short breaks from Facebook reduced anxiety, loneliness and negative social comparison. The documents form p… 11월 24, 2025