AI Safety news and analysis on Web Pulse: 23414 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Anthropic's public alignment work includes Petri, an open-source behavioral auditing tool, and Claude Opus 4.7 release notes, but these do not substantiate claims of specific safety improvements across 10 alignment failu…
AI Shipcheck, a new open-source static analysis tool released as v1.0.1, scans AI-generated code locally to flag production-readiness issues, scoring apps across nine categories and blocking deployment on critical findin…
Fox News reports that the AI infrastructure boom is creating a surge in blue-collar construction jobs, with firms scrambling to fill positions building data centers and power systems. OpenAI and over 100 companies warned…
A federal court in the Northern District of California ruled in favor of Anthropic, granting a permanent injunction and vacating Secretary of Defense Pete Hegseth's action that designated the AI company a national securi…
A revised estimate from Google Quantum AI, co-authored by Stanford professor Dan Boneh and Ethereum researcher Justin Drake, suggests that breaking the elliptic-curve cryptography securing Bitcoin and Ethereum may requir…
An investigation has found that Anthropic's Claude Code Auto Mode, which uses a safety classifier to autonomously execute commands, is vulnerable to indirect prompt injection attacks, achieving a 60-80% success rate in t…
Z.ai released the weights for its GLM-5.3 model on August 28, a 756 GB package with a broad commercial license, claiming steep gains in coding and vulnerability-hunting benchmarks. The model scored 28.3 on Terminal-Bench…
Cisco is partnering with Teleport to secure IT infrastructure access using short-lived cryptographic identities instead of static credentials, and is taking an undisclosed equity stake in the company. Teleport CEO Ev Kon…
A report from VentureBeat traces the rise of AI-related software incidents, citing that 45% of AI-generated code changes require debugging in production and the 2025 DORA report found a 10% increase in code instability l…
A developer studying for the AWS Certified AI Practitioner exam built a question-answering bot for the 2026 World Cup using Amazon Bedrock Knowledge Base and Guardrails, and found that the guardrail's denied topics and c…
UC San Diego researchers, led by doctoral student Sam Crow and cybersecurity expert Aaron Schulman, demonstrated a hack into a Boeing 737's flight management computer in under 60 seconds using a quarter-sized device, alt…
Multimodal large language models (LLMs) fail spectacularly on simple visual reasoning tasks due to a 'semantic drift' between visual embeddings and high-level concepts, according to a developer's testing. Even with struc…
The EU AI Act's high-risk system obligations were postponed by sixteen months via Regulation (EU) 2026/1744, which entered into force on 27 July 2026. The delay was due to missing harmonised standards and certification b…
A security researcher demonstrated that modern office AI agents can be hijacked via prompt injection, tricking the agent into executing malicious tool calls such as sending phishing emails to internal executives. The att…
Weir, a new open-source CI gate for AI agents, fails builds when sensitive data reaches a forbidden sink by reading OpenTelemetry traces the agent already emits. The tool, installable via `pip install weir-scan`, reconst…
URML, a new open-source safety-evaluation harness for AI agents operating lab and factory hardware, validates robot intent against declared device and deployment limits, issuing machine-readable refusals with evidence ta…
XAI, Elon Musk's artificial intelligence company, is suing users who generate deepfakes with its Grok chatbot, aiming to establish a legal precedent that shifts liability from the model to the user. The company argues th…
A US district judge permanently lifted the Pentagon's ban on Anthropic, ruling the decision to blacklist the Claude maker as a supply chain risk was "illegal and baseless" and that the national security justification was…
Ikanos, an open-source capability engine, introduces a security layer for MCP clients by replacing upstream server tool lists with reviewed, versioned YAML files, addressing vulnerabilities such as CVE-2025-54136 (MCPois…
OpenAI and more than 100 technology, cybersecurity, and financial organizations, including Google, Microsoft, Anthropic, AWS, Cloudflare, and CrowdStrike, signed an open letter calling for a global surge in cyber defense…