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.
Google DeepMind researchers published a framework for managing human-AI task delegation, identifying five core requirements including dynamic capability assessment and structural transparency. The paper treats delegation…
A bipartisan discussion draft from Rep. Lori Trahan (D-MA) and Rep. Jay Obernolte (R-CA) proposes a three-year federal preemption of state AI regulations while requiring top developers to disclose safety and security ris…
Independent security researcher Taylor Hornby discovered a critical soundness vulnerability in Zcash's Orchard shielded-pool circuit on May 29, 2026, while auditing the protocol for Shielded Labs, using Anthropic's Claud…
Site reliability engineers are rethinking the classic Four Golden Signals—latency, traffic, errors, and saturation—arguing these metrics fail to detect failures in AI systems where a service returns HTTP 200 but delivers…
Let’s Encrypt announced plans to implement post-quantum-safe web authentication using Merkle Tree Certificates (MTCs), aiming to protect TLS connections against future quantum computing threats. The organization targets …
Anthropic proposed a mechanism for pausing advanced AI development to allow human oversight and reflection. The company called for implementing a "pause button" that would enable society to assess risks before deploying …
A new report from the Center for Democracy & Technology (CDT) identifies dark patterns in AI chatbot design, including anthropomorphization and deceptive interfaces borrowed from social media, raising concerns about user…
A developer demonstrated that adding tools to an LLM application triples its attack surface, creating vectors for prompt injection, tool parameter injection, and information leakage. In comparative tests, a hardened agen…
Oinone, an open-source metadata-driven low-code framework, enables enterprises to use AI for coding without leaking sensitive data to third-party cloud models. The platform runs fully self-hosted and air-gapped, ensuring…
On June 3, 2026, attackers pushed malicious commits to the GitHub repository `icflorescu/mantine-datatable` and four sibling repos, planting a 4.3 MB payload from the Miasma worm family. The commit added configuration fi…
Anthropic published an analysis of 832 accounts banned for malicious cyber activity between March 2025 and March 2026, mapping their behavior to version 18 of the MITRE ATT&CK framework. The study found 67.3% of the acco…
OpenAI will comply with President Donald Trump's executive order requiring companies to provide federal officials access to AI models 30 days before public release for benchmarking, OpenAI Head of Countries George Osborn…
InfoQ published a six-article series titled "Securing the AI Stack: From Model to Production" on June 5, 2026, outlining a roadmap for deploying resilient AI systems. The series identifies data poisoning, AI-driven phish…
Erin Maus, a Unitarian Universalist software engineer, secured a religious exemption from using AI at her tech-entertainment job in mid-May after citing environmental and ethical objections. The accommodation comes as Po…
Senior US officials have held preliminary discussions with major AI companies about the federal government acquiring shares in firms building frontier AI, according to a report by NOTUS. OpenAI CEO Sam Altman has pitched…
OpenAI announced it will comply with President Trump's executive order mandating federal review of advanced artificial intelligence models before public release. The company stated it will submit its newest systems for g…
A survey of 272 AI experts from The University of Queensland and MIT FutureTech found a 10% probability of catastrophic AI outcomes within five years, including mass casualties, financial losses, or democratic collapse. …
NVIDIA at GTC 2026 presented the full-stack confidential computing capabilities of the Vera Rubin architecture, emphasizing that future AI chips must keep data encrypted throughout computation to prevent access by cloud …
A developer built a real-time activity logging and security scoring dashboard for Claude Code that tracks every file read, command executed, and API call with risk labels and timestamps. The tool, available as an npm pac…
Willow, an Israeli startup founded by three former Wix engineers, raised $7 million in a seed round led by Hetz Ventures to build an identity and access management platform for autonomous AI agents. The company’s platfor…