AI Safety news and analysis on Web Pulse: 13553 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Police in China are now patrolling with second-generation smart glasses that scan crowds, match faces to government databases, and identify individuals in milliseconds, with claimed accuracy rates between 95% and 99.5%. …
Microsoft warned that cybercriminals are using AI chatbot interactions and poisoned search results to direct users to malicious download sites in an active cryptojacking campaign. The campaign impersonates legitimate sof…
Enterprises relying on network policies, API gateways, and role-based access control (RBAC) for AI agent accountability face critical gaps, as these tools were designed for deterministic, human-driven workloads rather th…
OpenAI CEO Sam Altman acknowledged that the default personality of ChatGPT, used by 900 million people weekly, is the company's most impactful product decision yet lacks any scientific framework. Altman said he has enlis…
Cybersecurity researchers identified a malicious npm package named "mouse5212-super-formatter" designed to upload files from the "/mnt/user-data" directory used by Anthropic's Claude AI assistant. SOC Prime separately re…
A critical vulnerability tracked as CVE-2026-48710 and nicknamed "BadHost" was disclosed in the Starlette ASGI framework, affecting all versions prior to 1.0.1. The flaw allows attackers to manipulate the HTTP Host heade…
Full automation of AI research and development would likely produce a large acceleration in technological progress even without a software-only singularity, according to a new analysis. The AI Futures Model shows that in…
OpenAI is partnering with The Associated Press and Democracy Works to provide live vote counts and reliable voting information, aiming to combat AI-driven election misinformation ahead of upcoming elections in the US and…
Researchers have developed FuzzingBrain V2, a multi-agent large language model system that automatically discovers and reproduces software vulnerabilities. The system achieved a 90% detection rate on a competition datase…
AI coding agents are installing software packages from unverified sources that no organization claims ownership of, creating significant security and accountability risks in software development. The practice has raised …
Environmental activist Erin Brockovich launched a crowdsourced map tracking AI data center locations, which has already collected 2,716 public reports nationwide. The platform highlights how these facilities consume wate…
Pope Leo XIV, OpenAI CEO Sam Altman, and Nvidia CEO Jensen Huang have each recently offered measured perspectives on artificial intelligence, pushing back against both utopian and doomsday narratives. The pope's encyclic…
GCHQ director Anne Keast-Butler warned Wednesday that the U.K. and its allies face a "narrowing window" to stay ahead of Russia and China, as both nations escalate espionage and hybrid warfare against the West. Speaking …
California Governor Gavin Newsom is shifting toward a more populist stance on artificial intelligence regulation as he prepares for a potential 2028 presidential run, after years of aligning with Silicon Valley interests…
Amazon announced 70 recipients of its Amazon Research Awards from 49 universities across 11 countries, funding research in AI for information security, agentic AI, automated reasoning, cryptography, cybersecurity, and su…
A super PAC funded by OpenAI, Palantir, and a16z executives has spent millions attacking New York Assemblymember Alex Bores in the NY-12 Democratic primary. Bores raised over $2.2 million last quarter, including $364,000…
Prime Minister Mark Carney announced on May 27, 2026, that Canada's long-awaited federal AI strategy will be released next week. The plan, which missed earlier deadlines after consultations were fast-tracked, is expected…
More than half of organizations experienced an AI-related security incident or near miss last year, according to an Okta survey. Bosses remain overconfident about their awareness of shadow AI use by workers, leaving comp…
Researchers have identified brain-computer interfaces (BCIs) as a viable path to create moderately superintelligent humans, capable of outperforming current civilization's best minds in alignment and x-risk cognition by …
An Anthropic engineer has documented a critical safety practice for Claude Code sessions: immediately terminating any session that has executed a delete command. The engineer's analysis of a 70 MB session dump revealed t…