by Sandra Dziedzic | Mar 7, 2026 | AI News
A new study from King’s College London reveals that advanced AI systems often treat nuclear escalation as a rational strategic move, deploying tactical nuclear weapons in 95% of simulated geopolitical crises. The research highlights a concerning lack of “nuclear taboo” among frontier models, prompting experts to call for rigorous human oversight as governments increasingly integrate AI into defense and decision-making frameworks.
by Sandra Dziedzic | Mar 7, 2026 | AI News
Software engineer Sammy Azdoufal’s weekend experiment with a DJI Romo robot vacuum revealed a global smart-home security flaw affecting nearly 7,000 devices across 24 countries. By connecting the vacuum to a PlayStation 5 controller and using Claude Code to probe cloud communications, Azdoufal gained access to live camera feeds, microphone audio, and floor maps of homes.
The incident underscores the privacy risks of internet-connected devices: everyday smart-home tools can function as mobile data-gathering systems. While the vulnerability was responsibly disclosed and partially resolved by DJI, the episode highlights how convenience often outpaces security, raising urgent questions about cloud safety, device authentication, and consumer privacy.
by Sandra Dziedzic | Mar 7, 2026 | AI News
Pew Research Center has released a study examining AI use among American teenagers, revealing that the technology is fully integrated into daily life for research, coursework, and entertainment. Surveying 1,458 teens and parents, the study finds widespread adoption, alongside significant awareness of AI-assisted academic dishonesty — 60% of teens perceive cheating with AI as common, rising to 75% among active users.
Despite ethical concerns, teens largely view AI as a productivity enhancer, helping them “get things done faster” and making learning more efficient. The report also highlights a communication gap: four in ten parents have never discussed AI with their children, leaving peer-to-peer guidance to dominate.
The findings underscore a transitional moment: teenagers are growing up in an AI-native world, balancing increased capability with emerging risks in ethics, creativity, and future workforce readiness.
by Sandra Dziedzic | Mar 5, 2026 | AI News, OpenAI
Sam Altman has posted updates on X detailing significant revisions to OpenAI’s Pentagon contract following internal criticism, user cancellations, and a surge of sign-ups to Anthropic. The original deal, finalized rapidly after the Pentagon restricted Anthropic from federal use, drew criticism for being “opportunistic and sloppy,” according to Altman.
OpenAI clarified it will not deploy systems to the NSA or other DoD intelligence agencies while contract loopholes are addressed. The controversy underscores how military AI agreements can impact brand reputation, employee morale, and competitive positioning, with ethics in defense-related AI emerging as a key differentiator across the industry.
by Sandra Dziedzic | Mar 5, 2026 | AI News
MWC 2026 in Barcelona showcased the cutting edge of consumer tech, from robots to foldables. Highlights included Honor’s Robot Phone with a mechanical gimbal camera, Lenovo’s AI Workmate desktop robot and Work Companion smart display, and the ultra-thin Honor Magic V6 foldable at 8.75mm closed. Xiaomi debuted the Leica Leitzphone with a 1-inch sensor, while Tecno revealed a modular 4.9mm phone with ten swappable attachments.
In laptops and gaming, Lenovo introduced the Legion Go Fold handheld and Yoga Book Pro 3D concept, and Qualcomm unveiled its Snapdragon Wear Elite chip, powering AI on-device with unprecedented performance. Huawei previewed 6G-ready infrastructure with its U6GHz active antenna unit.
Three overarching trends emerged: on-device AI delivering real-time experiences, materials innovation such as silicon-carbon batteries and durable foldables, and ecosystem-first thinking linking phones, wearables, and PCs — with Samsung also pioneering AI-security integration via its KAI-to-Sentinel SIEM connector.
by Sandra Dziedzic | Mar 5, 2026 | AI News
Perplexity has unveiled Perplexity Computer, a multi-model orchestration system that distributes tasks across 19 different AI models, positioning the company among the first to make model flexibility a core feature of its platform. The details Users describe the...
