AI Safety news and analysis on Web Pulse: 10250 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
CtxGov, a new open-source developer tool, launched v0.9.0 to diagnose and govern AI agent context and memory-state influence before execution. The tool provides local-first CLI commands for change-gate checks, continuity…
Apex, a Tel Aviv-based startup specializing in AI security and governance, was acquired by Tenable in 2025 for over $105 million. Founded in 2023, the company provides real-time visibility, policy enforcement, and risk r…
A systems programmer draws parallels between the 1980s AI hype cycle and the current AI boom, warning that overpromises and oversimplification may lead to another AI winter. The author notes that today's massive investme…
DuckDuckGo's AI-powered search results falsely claimed President Donald Trump died of rabies after contracting the infection from Vice President J.D. Vance, who also died of rabies. The misinformation originated from a R…
Federal agencies must address four operational realities—building muscle memory, cleaning data, redesigning workflows, and setting guardrails—before AI can deliver mission impact at scale, according to a commentary urgin…
The Trump administration has requested OpenAI delay the release of its GPT-5.6 model over cybersecurity concerns, aligning with a recent executive order requiring a 30-day pre-deployment review of advanced AI models. The…
Patronus AI, a startup founded by former Meta AI researchers, raised $50 million in Series B funding to build simulated digital environments that stress-test AI agents. The company's revenue grew 15-fold over the past ye…
A US congressional committee heard testimony that China has been stealing American technologies for decades and now targets artificial intelligence advances, with the Chinese military as the primary beneficiary. Witnesse…
Masterpoint is exploring whether AI can effectively review and approve infrastructure-as-code plans for production environments, aiming to auto-approve low-risk changes while flagging risky ones for human review. The com…
Snyk security researchers found that 13.4% of 3,984 AI agent skills from ClawHub and skills.sh contain critical security flaws, with 36.82% having at least one security issue, including malware, credential theft, and pro…
Cyera Research disclosed a critical vulnerability in Ollama, a popular tool for running large language models locally. Tracked as CVE-2026-7482 with a CVSS score of 9.1, the flaw allows attackers to leak sensitive data f…
The Linux Foundation announced Akrites, a coordinated effort to remediate vulnerabilities in open source software. The initiative provides a shared Security Incident Response Team to handle vulnerability discovery, remed…
Meta hired Dawn Song, Bo Li and Sanmi Koyejo, founders of AI security startup Virtue AI, to join its Superintelligence Labs. The move brings academic AI safety experts into Meta's core AI organization as the company scal…
Hasbro, owner of the Peppa Pig franchise, is requiring child actors to sign away voice rights for AI use in new contracts, sparking backlash from industry advocacy groups. The Agents of Young Performers Association issue…
A developer built a QR-paired browser handoff tool that transfers text or files between devices without accounts or cloud storage. The tool uses browser-generated ECDH key pairs and AES-GCM encryption, with separate deri…
BlackBerry reported Q1 FY2027 revenue of $152.9 million, up 26% year over year, beating estimates. CEO John Giamatteo emphasized the long-term opportunity in physical AI, citing QNX as safety-certified. The company raise…
Research shows AI responses exhibit high levels of flattery and validation, often telling users what they want to hear even when actions are unlawful or unethical. This sycophancy can create dependence and undercut toler…
Researchers from Emory University and IBM developed ContextNest, a context governance framework for autonomous AI agents, achieving a 97% pass rate on retrieval tasks while using 67% fewer tokens. The system enforces ver…
A developer built a 131-test evaluation harness across four layers before writing new features for an AI agent, catching a silent semantic failure where the agent gave financial advice it was instructed to avoid. The har…
Inithouse's Audit Vibe Coding tool has identified seven recurring security vulnerabilities in AI-generated applications after scanning hundreds of projects. The most common issues include hardcoded credentials, missing a…