AI Safety news and analysis on Web Pulse: 11018 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
NVIDIA announced Halos for Robotics, the industry's first full-stack safety system for physical AI, extending its autonomous vehicle safety architecture to robotics. The system includes NVIDIA IGX Thor compute, Holoscan …
NVIDIA launched Halos for Robotics, a full-stack functional safety system for physical AI, extending its autonomous vehicle safety technology to industrial robots, humanoids, and autonomous mobile robots. The platform co…
GLM-5.2, a new open-source AI model, has been released, boasting strong benchmarks and positioning as the best open model to date, though it remains behind the frontier by 4-7 months. The model is a significant step up f…
Mozilla distinguished engineer Brian Grinstead and his team used an agentic bug-finding pipeline powered by Anthropic's Claude Mythos model to discover a 15-year-old bug in Firefox, contributing to a record month of secu…
Skadden attorneys warn that using AI to record or transcribe board meetings and generate minutes creates detailed records of confidential discussions that could be turned over in litigation. Communications between direct…
AI tools remove friction from research tasks, but that friction also served as an invisible quality gate where errors were caught. As toil disappears, so does the examination that rode on it, leaving users unaware of wha…
Data poisoning attacks are a growing threat to machine learning pipelines, where adversaries subtly manipulate training data to compromise model behavior. Attackers use diverse techniques to inject malicious examples tha…
CIOs identify scaling AI for business value and cybersecurity as the top challenges in 2026, with 71% struggling to prioritize AI use cases and 59% of AI initiatives failing to reach production, while AI-driven attacks e…
HTPBE, a structural forensics tool, distinguishes between content classification and structural analysis for detecting AI-generated PDFs. It identifies documents rendered by headless browsers, Python libraries, or other …
Cory Doctorow argues in a new book and podcast interview that the AI boom is driven by unsustainable hype and power dynamics that benefit bosses at the expense of workers, coining the term "enshittification" to describe …
A new phenomenon called 'Vibe Management' is emerging where middle managers use AI like Claude to generate strategy documents without genuine thought, creating a digital echo chamber of hallucinated strategies. This pose…
AI-generated websites, deepfake videos, and convincing phishing campaigns are making World Cup scams harder to spot ahead of the 2026 FIFA World Cup. Over 13,000 FIFA-themed domains were registered between January and Ma…
Cisco, Equinix, and Nvidia expanded their partnership on June 16 to deploy Cisco's Secure AI Factory across Equinix's global network of over 280 data centers, aiming to standardize enterprise AI deployments and bridge th…
Researchers trained deep MLPs to memorize secret binary strings and found that naturally learned behaviors resist detection by both human-made input-optimization algorithms and learned algorithms, posing a challenge for …
Three recent AI events reveal the fragility of relying on third-party model APIs. Anthropic suspended access to its Fable 5 and Mythos 5 models after a US government export control directive, causing a global blackout. M…
A developer with over a year of experience using LLMs and agents for coding warns that 'vibe coding' creates a false sense of productivity. The shift from writing code to reviewing AI-generated code increases cognitive l…
A Singaporean man, identified only as Luke, filed a police report after discovering a former schoolmate used AI to generate fake photos depicting them as a family, including images of a pregnancy and newborn. The man, wh…
Twenty-three code-executing plugins on ClawHub, an AI agent registry, were found squatting under official @openclaw and @clawhub scopes because those scopes were not reserved for their owners, exposing security gaps in A…
Mobilintec, a mobility startup, will develop an agentic AI-based traffic collision prediction and response system for local roads as part of South Korea's 2026 AI City project, funded by the Ministry of Land, Infrastruct…
Europol officials estimate that 90% of online content will be AI-generated by the end of 2026, up from 50% currently driven by bots and AI. Google, which accounts for 90% of global searches, is criticized for prioritizin…