AI Safety news and analysis on Web Pulse: 13110 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
OWASP released AISVS 1.0, a security checklist with 514 verifiable requirements for AI systems, at Global AppSec Vienna. The standard fills a gap by providing testable controls for prompt injection, MCP security, and sup…
AI-powered vulnerability discovery tools are driving a surge in disclosed software flaws, with the Forum of Incident Response and Security Teams (FIRST) now projecting 66,000 CVEs in 2026, far exceeding initial estimates…
US export controls on advanced AI models have prompted Indian venture capitalists and AI leaders to call for accelerated domestic AI capability. Anthropic disabled its Fable 5 and Mythos 5 models for foreign customers to…
A New York Post opinion piece published June 15, 2026, by developmental scientists and educators argues that AI-powered toys are displacing traditional play and raising developmental concerns. The authors cite a study sh…
A group of cybersecurity experts led by former Facebook chief security officer Alex Stamos is urging the US government to lift restrictions on non-US nationals' access to Anthropic's frontier models Mythos 5 and Fable 5,…
Zcash surged 25% after founder Zooko Wilcox confirmed that Anthropic AI's Mythos model completed a security audit finding no new critical vulnerabilities, reversing a 40% drop from a prior forgery bug. The audit by the p…
A tech professional warns that AI has already passed a tipping point with tools like Fable and Opus, enabling rapid translation of vague ideas into shippable products, and predicts brutal economic disruption as companies…
A user named Zhynx posted on X that AI systems could become obsolete quickly, citing Claude Fable's three-day lifespan and noting that open-weight models are only months behind the frontier, with decentralization and ope…
Coding agent sandboxes address host compromise but not credential authorization, leaving agents with high-scope tokens that can push to protected branches, deploy infrastructure, or read secrets. The real risk is authori…
A new guide for cloud architects outlines practical steps to govern AI in the cloud, addressing the growing threat of shadow AI. The guide emphasizes discovery, data classification, and policy-as-code to secure AI deploy…
Apple announced expanded parental controls in macOS 27 and iOS 27 at WWDC 2026, including redesigned Screen Time, contact approval, website permissions, and age-based protections. These features aim to make the MacBook N…
AI safety is a software architecture problem that requires treating AI systems as production systems with proper trust boundaries, permissions, and logging. Developers must avoid making the model the policy engine and in…
The widespread adoption of AI, regardless of its actual efficacy, poses a significant threat to the labor force by enabling de-skilling and mass layoffs, as corporations use AI as a pretext to replace human workers and l…
Silicon UK's AI For Your Business podcast explores autonomous money and AI spending in an episode titled 'Autonomous Money: Are We Ready to Let AI Spend for Us?', examining accountability, audit trails, and governance ch…
Anthropic received a U.S. export-control directive requiring suspension of access to its Fable 5 and Mythos 5 models for foreign nationals, citing national-security concerns. The directive, issued without detailed justif…
AI-generated content, or 'slop,' is shifting cultural norms by exploiting human preferences for symmetry, legibility, and visceral appeal, a process the author calls 'hyperslopification.' This hyperpalatable, low-cost co…
The Red Team AI Benchmark v1.9.0 introduces a modular scoring architecture, unified provider interface, and YAML-native configuration to improve usability for legitimate researchers. The update also adds an ethical use p…
NexArt introduces Certified Execution Records (CERs) for AI systems, enabling independent verification of execution records without trusting the provider. The CER uses canonical JSON, integrity anchoring, and signed trus…
The industry-standard code review process, review-then-commit, is broken by AI coding agents, creating a principal-agent problem where contributors generate low-effort pull requests that overload reviewers. This imbalanc…
The European Commission is assessing the implications of a US order that forced Anthropic to shut down its Fable 5 and Mythos 5 models worldwide, sparking a debate across Europe about AI sovereignty. European researchers…